Corporate social responsibility and export performance under stakeholder view: The mediation of innovation and the moderation of the legal form

Prior confronting findings on the association between corporate social responsibility and profitability show that some aspects remain to be clarified about this binomial. Contributing to this body of research is particularly interesting in the case of the exporters to understand the role that corporate social responsibility may play in improving export performance. This study aims to determine the impact of firms' corporate social responsibility efforts on their export performance by examining innovation and the legal form of the organization as factors that could affect this relationship. Using data from 107 agri-food companies, we found that corporate social responsibility has no direct effect on export performance. However, our results show that innovation serves as a mediator in this relationship. It was also found that companies that adopt associative legal forms (i

and, thus, to be perceived as more attractive entrants than less CSRoriented firms (Khojastehpour & Saleh, 2019;Laplume et al., 2008).This fact exerts pressure to engage in CSR practices, as CSR would contribute to export firm's and, a priori, to enhance export performance (EXP), typically defined as the extent to which firms achieve their economic (e.g., profits, sales) and strategic (e.g., market expansion) objectives regarding exporting products abroad (Cavusgil & Zou, 1994).
To be competitive in the international context, there is an increasing number of companies that are adopting differentiation approaches based on CSR as the essential element of this type of competitive strategies.For instance, La Palma, a Spanish fruit and vegetable cooperative that exports 90% of its production, has recently been awarded internationally with the 'Fruit Logistica Innovation Gold Award 2022' for its innovation strategy focused on achieving CSR practices namely through 'less is more' agriculture, that is, less water, less CO2, less waste, more performance, and more productivity (La Palma, 2022).This example illustrates the importance that CSR orientation arises in the export context.However, this is a topic that has received scarce attention from researchers to date.In fact, research on Stakeholder Theory has suggested that CSR is positively related to firm performance (e.g., Laplume et al., 2008;Peloza & Shang, 2011), but the relationship between CSR and EXP has been scarcely addressed in the academic literature (Kim, 2016;Xu et al., 2018).Previous studies that examine this association in different sectors have showed controversial results.Some have argued that there is a positive relationship between CSR practices and EXP (e.g., Antonietti & Marzucchi, 2014;Boehe & Barin Cruz, 2010;Xu et al., 2018).For instance, Boehe and Barin Cruz (2010) found a positive and significant effect of CSR product differentiation on EXP.Similarly, drawing upon firm heterogeneity theory, Antonietti and Marzucchi (2014) showed that investment in green strategies could increase productivity, which leads to a higher probability of exporting in international markets with more stringent environmental regulation.However, other studies have found a negative association (Costa et al., 2015;Kim, 2016).For instance, relying on a sample of 438 firms listed on the Korean stock market, Kim (2016) found that increases in CSR scores leaded to a decrease in export sales.Similarly, Costa et al. (2015) found that, in the case of technology exporters, the adoption of CSR practices decreases the positive effect of exploitative innovation on EXP.
These confronting results highlight the need to probe further into the analysis of this association (Costa et al., 2015;Xu et al., 2018), specially analyzing the mechanisms that underlie the connection between the adoption of CSR practices and higher EXP (Norheim-Hansen, 2015).This study delves further into the CSR-EXP relationship, examining the role that innovation (IN) and the legal form of organization (hereinafter, LFO) play in it.On the one hand, the adoption of CSR practices could contribute to improve the IN of companies insofar as it creates more IN-friendly working environments and promotes business cooperation on IN (Rexhepi et al., 2013).IN enables companies to meet the commitments made in the process of dialog and cooperation with their stakeholders (Darnall et al., 2010) and thus provide business solutions to certain social and environmental challenges (Gallardo-Vázquez & Sánchez-Hernández, 2013;McWilliams & Siegel, 2000;Surroca et al., 2010).Further, in an international business scope, IN could be an essential intangible resource that confers competitive advantage, based on differentiation, to companies operating in foreign markets, while allowing them to adapt to changing environments ahead of their competitors (Azar & Ciabuschi, 2017;Damanpour et al., 2009).This competitive advantage based on IN can be translated into an increase in EXP (Lefebvre et al., 1998;Pla-Barber & Alegre, 2007).Combining both relationships, this work proposes that there may be a mediating effect of IN on the link between CSR and EXP.Testing this mediation effect could provide a novelty to the existing literature as it may shed light on the underlying mechanisms on the relationship between CSR and EXP.
On the other hand, when studying the relationship between the adoption of CSR practices and performance, it is essential to consider whether the social conception is present in the firm's legal form (Castilla-Polo et al., 2018).A company whose legal form is more democratic and cooperative tends to be more CSR-oriented (Lund-Thomsen & Lindgreen, 2014) and this socially and environmentally responsible orientation may influence on its performance.At this point, a distinction is made between associative companies (i.e., cooperatives) and non-associative companies (i.e., capital-focused companies).Previous studies have shown the close relationship between the pairing CSR-cooperativism and performance (Castilla-Polo et al., 2018;Castilla-Polo & Sánchez-Hernández, 2020;Forcadell, 2005).For example, Castilla-Polo et al. (2018) showed that there is a positive relationship between the adoption of CSR practices and cooperativism and this, in turn, was associated with higher profitability in cooperative companies.These previous results invite further empirical study regarding whether the LFO can moderate the potential effect of CSR on IN and EXP, in such a way that associative LFOs may intensify these relationships.
Therefore, the aim of this work is threefold.First, to analyze to what extent CSR has a positive impact on the EXP of companies, checking whether this relationship is direct or, on the contrary, requires mediation mechanisms.Second, to understand the role played by IN in the relationship between CSR and EXP, analyzing whether it acts as a mediating variable in this association.Finally, to check whether the LFO of companies affects the intensity of the relationship between CSR and IN and EXP.
To empirically test these questions, a sample of 107 companies in the agri-food sector in Almería, in southeast Spain, was analyzed.The agri-food sector in this area of Spain has served as the research context in many studies on CSR (e.g., Briones-Peñalver et al., 2018;Galdeano-G omez et al., 2017).However, the CSR-EXP relationship has rarely been examined in the agri-food sector, despite its strategic importance as a main pillar of the export sector in many economies such as Spain.For instance, in 2021, 16.55% of the Spanish total goods exported came from the agri-food sector (Spanish Ministry of Agriculture, Fisheries and Food, 2022).Further, Almería is one of the geographical areas with the highest business concentration in the agri-food sector in Europe.Consequently, since our sample contains companies which operate in this highly competitive export context, our results may expand previous research in this setting and shed additional light on the association between CSR and EXP.This paper contributes to previous literature that calls for further research on the mechanisms that affect the relationship between CSR and performance (Costa et al., 2015;Kim, 2016;Laplume et al., 2008;Peloza & Shang, 2011).Empirical evidence is provided of the existence of a mediation of IN in the CSR-EXP relationship, as well as of the existence of a close link between CSR and IN that is significantly stronger in the case of companies with associative LFOs, confirming the moderating effect of the LFO on the relationship between CSR and IN.

| THEORETICAL FRAMEWORK: CSR UNDER THE STAKEHOLDER THEORY IN EXPORT CONTEXTS
CSR is a concept that is constantly being revised (Carroll, 2021;Matten & Moon, 2008, 2020), and one that has been adapted to the economic, political, and social reality of each moment (Sarkar & Searcy, 2016).There are numerous academic definitions and theoretical perspectives around this concept (Matten & Moon, 2008), which have generated some confusion in companies when it comes to applying them to their own management situation (Carroll, 1999).
Nevertheless, in the multiple definitions of CSR, there is a common element which is that CSR is understood as the voluntary undertaking on the part of the company of several responsibilities ( Van, 2003) which go beyond legal and regulatory requirements (McWilliams & Siegel, 2001).These responsibilities translate into a long-term commitment to incorporate in their management decisions the expectations and interests of all agents involved in or affected by the company's activity (Smith, 2003).CSR implies a strategic and proactive way of doing business, beyond social action, which must consider economic, environmental, and social aspects in a balanced way (Elkington, 1998;Panwar et al., 2010), and must be understood not as an expense, but as an investment that can improve the company's competitiveness (Inyang et al., 2011).CSR is therefore defined as the voluntary, proactive, and strategic commitment of the company to meet the expectations of its stakeholders and to preserve the environment, beyond legal requirements, and which development and application can create value in the long term.This vision of the CSR concept is intimately related to the Stakeholder Theory (Donaldson & Preston, 1995;Freeman, 1984), which is one of the most used approaches to conceptualize and explain CSR and it is accepted by most of the institutions that regulate or promote it (Carroll, 2021).This theory presents the company as an organization composed of a plurality of agents (stakeholders) who affect or are affected by its management decisions (Freeman, 1984).Consequently, companies must recognize the interests of these agents, as well as the existence of a network of relationships and reciprocal expectations of behavior between them, facilitating their participation in the business decisions that affect them directly (Evan & Freeman, 1988;Freeman, 1984).From this theoretical perspective, CSR could be formed by different components referred to the company's level of commitment to their stakeholders.
Drawing upon the Stakeholder Theory in the international business context, the theoretical reasoning might lead to think that a higher level of commitment to stakeholders (including employees, farmers, customers), lead to better results in terms of exportations (Arora & De, 2020).However, this association is not so obvious in the international agri-food sector.In fact, prior literature has suggested that, in firms belonging to primary sectors (e.g., agriculture, farming), the stakeholders' influence in adopting CSR practices may not be decisive (Briones-Peñalver et al., 2018;Guoyou et al., 2013;Marshall et al., 2010).For instance, using a sample of export firms in the wine industry in New Zealand and the United States, Marshall et al. (2010) found that neither the internal stakeholders nor the external stakeholders played a significant role in influencing the adoption of CSR practices.
This work considers that, to analyze the demands of stakeholders in an export context, it is essential to theoretically differentiate between two perspectives: the international stakeholders' point of view and the local stakeholders' perspective.Related to the international stakeholders' perspective, final consumers in the host country do not usually know or recognize the farmers in the country of origin of the product.Further, in this sector firms do not usually maintain stable export networks and instead use intermediary's companies for selling their products.As a result, there may be scarce (or null) interaction or feedback from the final customer in the host country, even in the case of products with ecolabelling or other type of product certification (Arora & De, 2020).In this scenario, in which part of the stakeholders (e.g., final consumers) do not know who is the producer, how could a social and environmentally responsible firm (at home) leverage differentiation abroad in terms of brand image or better positioning with its stakeholders?In the international business contexts, a firm that is well recognized as CSR-focused by key stakeholders (Arora & De, 2020;Khan et al., 2013) such as intermediaries and corporate clients in the agri-food sector, could obtain greater benefits in comparison with non-CSR-focused firms.In fact, companies show more interest in adopting CSR practices when they participate in the export market, which is positively reflected in their EXP (Arora & De, 2020).
At host countries, intermediaries' expectations are related to information transparency and the availability of a wide range of quality, reliable products and services provided by export firms, and the adoption of CSR practices tend to favor these qualities.Thus, a higher CSR approach of the exporting firm could be translated as greater appreciation by the intermediaries (Barin Cruz et al., 2015;Boehe & Barin Cruz, 2010).This positive reputational effect due to the interest of intermediaries in acquiring products from the most responsible companies in the place of origin (Motlaghi & Mostafavi, 2017) can be transferred to the rest of the stakeholders, such as final consumers, through a greater presence in foreign markets.For instance, in the case of the agri-food sector, an international supermarket chain will be more interested in purchasing products from companies that adopt CSR practices than from non-CSR-adopter companies.Based on this argumentation, we consider that, when there is a greater presence of products from socially and environmentally responsible companies (at home) on the shelves of markets abroad, the positive effect of satisfying the demands of one group of stakeholders (e.g., intermediaries) will spread to other stakeholders' groups (e.g., end consumers).
Concerning the local stakeholders' perspective, several benefits of being a CSR-focused firm exist.For instance, socially and environmentally responsible companies are more likely to satisfy demands of farmers, whose expectations are mainly focused on improving the prices settled for their products and receiving technical and agronomic advisory services (Briones-Peñalver et al., 2018).Similarly, local employees' expectations are related to work-life balance, fair pay, job stability, among other labor aspects, which are often better accomplished by CSR-oriented firms (Aguinis et al., 2020;Forcadell, 2005).
In the same line, the activity of local agri-food companies generates significant negative externalities in the community, while the adoption of CSR practices can contribute to the viability of rural areas and a balanced development from a territorial point of view (Briones-Peñalver et al., 2018).
In sum, considering both international and local stakeholders' perspectives, acting responsibly could lead to meet all stakeholders' expectations and, thus to better long-term results for the firm (Godos-Díez et al., 2018).This reasoning is similar to Murillo-Luna et al. (2008) who argued that, in considering stakeholders' demands concerning CSR, one single dimension exists, and both stakeholders' perspectives (international and local) finally converge in a positive relationship between the adoption of CSR practices and the satisfaction of the stakeholders' interests.

| Corporate social responsibility and export performance
The relationship between the adoption of CSR practices and EXP has been studied in prior research (Antonietti & Marzucchi, 2014;Boehe & Barin Cruz, 2010;Costa et al., 2015;Kim, 2016;Xu et al., 2018).For example, Xu et al. (2018) argued that firms with high level of energy conservation and emission reduction are more likely to pursue international expansion because of the signaling effect of international certification, and this has a significant effect on EXP.Boehe and Barin Cruz (2010) also claimed that differentiation based on CSR is a more important predictor for EXP than quality-based differentiation.In the same line, Siegel and Vitaliano (2007) argued that CSR could be integrated as a fundamental element in a competitive strategy based on differentiation.Similar, concerning Stakeholder Theory, a key element in this differentiation strategy is the recognition of companies as being socially and environmentally responsible by its salient stakeholders because of the social and environmental actions these companies carry out (Barin Cruz et al., 2015;Darnall et al., 2010).
Nevertheless, in the international business context, the process of adapting to the expectations of stakeholders is more complex when companies operate in cross-border markets (Khojastehpour & Shams, 2020) due to, at least, two main reasons.On the one hand, companies that operate in international markets must face higher social and environmental standards.Companies must adapt to cultural norms and constraints and to the perception that international stakeholders have of the CSR (Matten & Moon, 2008).Indeed, multiple stakeholders could pose competing demands (Costa et al., 2015;Murillo-Luna et al., 2008).For instance, in some cases foreign stakeholders may be more sensitive to the ecological characteristics of products than local stakeholders (Guoyou et al., 2013;Khojastehpour & Shams, 2020).On the other hand, in some industries, such as international agri-food sector, final consumers can rarely connect with producers (Barin Cruz et al., 2015).At this point, it is important to note that the pressure that final consumers exert on commercial intermediaries makes these intermediaries add to their responsibilities (e.g., supply guarantee, quality commitment, fair prices) the CSR needs and expectations of this group (Khan et al., 2013).Consequently, intermediaries make their own these needs and thus demand differentiated products and processes that fulfill final consumers expectations.This aspect of the differentiation strategy can lead to an increase in the volume and performance of exports, as it helps to increase the perceived value of products, strengthen brand image, and increase the loyalty of international customers (Barin Cruz et al., 2015;Motlaghi & Mostafavi, 2017).As such, the adoption of CSR practices can help international companies to differentiate themselves from the rest of their competitors in foreign markets and, consequently, improve their EXP (Barin Cruz et al., 2015;Boehe & Barin Cruz, 2010;Motlaghi & Mostafavi, 2017).For instance, Khan et al. (2013) showed that Bangladeshi companies in export-oriented industries disclosure significantly more CSR practices, and the pressure exerted by important stakeholder groups was the most potent factor for CSR disclosures (Khan et al., 2013).
However, not all results show a positive and statistically significant relationship between CSR and EXP and scholars have argued that often stakeholders (e.g., customers, retailers) do not pay much interest on CSR issues, do not identify the CSR differentiation, or simply view compliance with CSR principles as the minimum level to operate (Boehe & Barin Cruz, 2010;McWilliams & Siegel, 2000).For instance, Costa et al. (2015) found that CSR practices have a detrimental impact on EXP through the effect of exploitative IN (compared to exploratory IN) and the effect of CSR on building stronger exploitative IN was not significant.Hence, the impact of the export firm's efforts to incorporate CSR practices may have no effect on increasing EXP.These confronting results encourage further study of the relationship between CSR and EXP.
Based on the above arguments, we postulate a positive CSR-EXP relationship as there are CSR elements often related to how companies must respond to the demands of their stakeholders, such as social or environmental certifications (Xu et al., 2018) or to have a certain positive reputation in terms of CSR (Barin Cruz et al., 2015), that allow firms to assume a privileged position in the market.

Hence:
H1.: There is a positive relationship between CSR and EXP.

| Corporate social responsibility and innovation
There are three ways in which CSR can support both technological and organizational IN (European Competitiveness Report, 2008;OECD, 2005).First, IN can be the result of a company's commitment to its stakeholders.CSR requires dialog and cooperation with stakeholders, which sometimes involves making a series of commitments to them that may encompass the creation of new products or services, or the improvement of existing ones, as well as certain organizational changes (Rexhepi et al., 2013).This dialog with the stakeholders, to consider their CSR interests, is even more necessary in the case of exporting companies that attempt to satisfy the demands of both local and foreign stakeholders (Guoyou et al., 2013;Khojastehpour & Shams, 2020).Secondly, through IN, the company can explore new business opportunities related to certain social and environmental needs such as poverty, inequality, social inclusion, the aging population, or climate change (Ratajczak & Szutowski, 2016).Finally, CSR also tends to create more innovative working environments by fostering employee trust, participation, and creativity, while promoting IN cooperation with other companies through agreements that encourage knowledge sharing in the design of new products and services (Rexhepi et al., 2013).Thus, CSR can become a powerful driver of IN (European Competitiveness Report, 2008).In fact, previous studies have concluded that IN increases when it becomes more CSRoriented (e.g., Gallardo-Vázquez & Sánchez-Hernández, 2013;McWilliams & Siegel, 2000).Some of them consider CSR as an IN itself (Surroca et al., 2010) since its implementation requires the application of certain organizational IN as well as significant strategic changes (Costa et al., 2015).
Drawing on Stakeholder Theory, several studies have demonstrated that greater IN could be positively related to the adoption of CSR practices in SMEs (e.g., Darnall et al., 2010;Esty & Winston, 2006).For instance, Darnall et al. (2010) argued that the stronger IN propensity is fueled by the smaller firm's interest in protecting the natural environment that is triggered by stakeholder concerns, specially perceived pressures from value chain, internal, and regulatory stakeholders.From these arguments, we propose the following hypothesis: H2.: There is a positive relationship between CSR and IN.

| Innovation and export performance
IN is a crucial element in the growth strategy of companies (Gunday et al., 2011).It is an intangible resource, which, by its very nature and characteristics, is difficult to imitate (Monreal-Pérez et al., 2012) (Boehe & Barin Cruz, 2010;Costa et al., 2015).
Expanding their activities to other markets allows companies to exploit their INs and thereby improve their performance, increasing their sales volume and expanding their potential customer base (Leonidou et al., 2007).Prior studies have concluded that technological INs increase the propensity to export and are an important determinant of EXP (Lefebvre et al., 1998).Other studies also conclude that organizational INs improve export performance both directly and indirectly by facilitating the efficient use of technological INs (e.g., Azar & Ciabuschi, 2017).Accordingly, we propose that: H3.: There is a positive relationship between the IN and EXP.

| The mediation effect of IN on the CSR-EXP relationship
Several studies have examined the critical role that certain mediating variables such as corporate reputation (Saeidi et al., 2015), customer satisfaction (Ali et al., 2020) or IN (Anser et al., 2018;Briones-Peñalver et al., 2018), can play in the relationship between CSR and performance.However, according to Nguyen and Nguyen (2020), only 23% of the studies, carried out between 2015 and 2020, which explore this relationship have included mediating variables in their models.
Several arguments lead us to believe that IN has a mediating role in the relationship between CSR orientation and export performance.
In the international business context, meeting the expectations of foreign stakeholders in markets with high social and environmental standards requires companies to make significant efforts in exploratory IN (Costa et al., 2015).In addition, these efforts in IN for satisfying international demands are usually related to the adoption of CSR-related practices (Boehe & Barin Cruz, 2010;Costa et al., 2015).For example, foreign customers play an important role in companies' adoption of green IN strategies in product and process (Guoyou et al., 2013).
The constant and proactive dialog with stakeholders, which is usually inherent to the adoption of CSR-practices (Darnall et al., 2010;Khojastehpour & Shams, 2020;Rexhepi et al., 2013), generates the opportunity to convert the needs of certain stakeholders into innovative actions, allowing companies to increase their global competitiveness (Guadamillas & Donate, 2011) and, consequently, to enhance their performance (García-Castro et al., 2010;De la Cuesta González, 2004).In this process, firms that adopt CSR practices as an important part of the differentiation strategy will need to incorporate R&D investment, which may result in both CSR-related process and product INs (McWilliams & Siegel, 2001).In this regard, when companies use the adoption of CSR practices to differentiate products abroad (Ali et al., 2020;Barin Cruz et al., 2015), they are likely to derive greater benefits from the resulting reputation building or maintenance (see McWilliams et al., 2006).As the inclusion of social and environmental issues in a company requires the adoption of INs (McWilliams & Siegel, 2001), IN may represent one of the missing links that explains the relationship between CSR and performance (Bocquet et al., 2017).Likewise, CSR supposes a modification of the relationships between a firm and their stakeholders, which implies changes at the management and organizational levels (Rexhepi et al., 2013) that can even be seen INs in themselves (Surroca et al., 2010).Therefore, it could be concluded that IN is the factor that may facilitate the differentiation strategy based on a CSR orientation, the competitive advantage derived from it, and the consequent increase in EXP.
Previous studies suggest the importance of including IN as a mediating variable in theoretical models that analyze the relationship between CSR and profitability (Bocquet et al., 2017;McWilliams & Siegel, 2000;Surroca et al., 2010).For example, McWilliams and Siegel (2000) found that CSR does not have a direct but an indirect effect on financial performance, through investment in R&D.On the other hand, Surroca et al. (2010)

| The moderating effect of LFO on the CSR-EXP and CSR-IN relationships
The type of legal form of the firm could affect the effect of the adoption of CSR practices (Castilla-Polo et al., 2018) because there are some LFO, such as cooperatives, that are focused on social aspects in a greater extent than capital-focused LFO (Forcadell, 2005).Cooperatives, as an LFO, are societies which are characterized by their markedly democratic nature (i.e., each partner has one vote, regardless of the capital contributed) and whose results are distributed in proportion to the productive effort contributed by each partner (International Cooperative Alliance, 1995).In capital companies, however, decisions are weighted according to the amount contributed by the partner and the results are distributed according to the participations of each partner.Cooperatives are based on values which are translated into a series of principles and operational rules for economic and organizational functioning, the so-called "cooperative principles" (International Cooperative Alliance, 1995) such as "management, participation and democratic control by the partners", "cooperation with other cooperatives", "commitment to the community", "education, training and information", and so forth, which are closely to the CSR concept in its different manifestations.In this sense, cooperative societies share the "cooperative principles" as a part of their business culture and this fact could encourage and facilitate the adoption of CSR practices (Castilla-Polo & Sánchez-Hernández, 2020).For instance, "in a democratic organization, establishing a mission founded on democratic values is key, so that a shared understanding can be developed among the members (and the rest of the stakeholders) about the raison d'être and the goals of the organization" (Forcadell, 2005, p. 269).These principles make CSR intrinsic to their very nature (Moral & Poyatos, 2010) and companies tend to benefit most from CSR practices when their adoption is based on internal organizational aspects (Tang et al., 2012).The application of certain principles such as self-help, self-responsibility, democracy, equality, equity, and solidarity, can make their involvement with CSR practices more systematic, regular, and consistent.This strengthens the confidence the stakeholders have in the company's commitment to CSR not being random or opportunistic, nor responding to occasional external pressure (Frooman, 1999).This greater consistency in the companies' commitment to CSR increases the intensity of the relationship between CSR and firm performance (Tang et al., 2012).
Related to EXP, cooperatives that operate internationally have flexible manufacturing systems for adapting to multiple markets and customer requirements and complying with safety regulations and approval regimes from several countries (Forcadell, 2005).Consequently, adopting cooperative as a legal form could make the involvement of the adoption of CSR practices more consistent (Moral & Poyatos, 2010;Tang et al., 2012) which can, in turn, make the relationship between CSR and EXP more intense.According to these arguments, we hypothesized: H5a.: The LFO moderates the relationship between CSR and EXP, making it more intense in the case of cooperative legal forms.
On the other hand, the importance that cooperative societies give to training and cooperation encourages IN (Forcadell, 2005;Garcés-Ayerbe et al., 2019), insofar as it favors their capacity to identify, assimilate, transform, and exploit new external knowledge (Garcés-Ayerbe et al., 2019;Marcuello & Saz, 2008), as well as to develop business networks, local production systems or industrial clusters (Bakaikoa et al., 2004).For instance, to foster eco-innovation process within companies, Garcés-Ayerbe et al. ( 2019

| Sample and data collection
The empirical study was developed through a survey aimed at export companies in the context of the agri-food sector in Almería, southeast of Spain.Analyzing the adoption of CSR practices in the international agri-food sector in this area of Spain is especially relevant for at least two reasons.On the one hand, its important economic and social contribution have determined the viability of rural areas.In the case of Almería, agri-food is a sector based on a network of small farms and family businesses, with a clear commercial vocation towards foreign markets and a high capacity for IN.In fact, in the 2020/2021 season, 80% of total production was exported (Cajamar, 2021).The agri-food sector has allowed Almería to converge in per capita income with the Spanish average and to be at the forefront of Andalusia, the region in which Almería is located.For instance, in 2019 the per capita income of Almería represents 80.2% of the Spanish average and 108.5% of the Andalusian average (National Statistics Institute, 2020).In fact, a large proportion of the Almería population is directly or indirectly related to this sector.On the other hand, in this geographical region, the agri-food sector has implied a significant improvement concerning environmental impact (Briones-Peñalver et al., 2018;Galdeano-G omez et al., 2017).This positive environmental effect has been achieved by implementing important changes in technologies and agricultural practices to reduce negative environmental impacts.For example, in the 2020/2021 season, the area cultivated under biological control techniques represented 40.6% of the total cultivated area in the province (Cajamar, 2021).In sum, the international and environmental focus of agri-food sector in Almería has resulted in significant contributions to the economic, social, and environmental spheres.
The sample was selected from the total population of trading companies in the agri-food sector.According to the information extracted from the SABI (Iberian Balance Sheet Analysis System) database, the agri-food sector in Almería is made up of a total population of 287 companies.Most of them are SMEs with an average annual turnover of 14.8 million euros and an average number of employees of 69.In order for the companies in the sample to represent a high percentage of the sector's turnover and employment and to have a minimum organizational infrastructure, a subset of 164 companies with a turnover higher than 5 million euros was selected.A selfadministered online questionnaire was sent to these companies, obtaining a total of 107 valid responses, 61% of which were answered by the company's president or manager, 25% by the human resources managers and the rest by other company managers.The sample size exceeds the minimum threshold necessary for the application of the technique used (Partial Least Squares modeling, hereinafter PLS) in this research and to reach acceptable levels of statistical power (Reinartz et al., 2009).
To avoid potential psychological biases linked to subjective measurement, we carried out several actions following the recommendations of different scholars (e.g., Skjong & Wentworth, 2001).First, a self-administered survey was used to ensure the anonymity of respondents and to avoid the tendency to give socially acceptable answers.
Second, efforts were made to ensure that the respondent had a broad and cross-sectional knowledge of the business.Third, we also ensured that variables and dimensions were clearly defined in the questionnaire and that the items represented them clearly and concisely.

| Measures
A list of items was sourced from the existing literature and adapted to the context of this study.According to the guidelines established by Skjong and Wentworth (2001), these items were subsequently reviewed by a panel composed of 13 experts.Table A1 shows the items used for measuring each construct.

| Dependent variable: export performance
Researchers in the field of international business have reached a consensus concluding that export performance is a broad construct composed of not only "increase in sales", but also customer satisfaction, among other variables (Lages et al., 2008;Tsinopoulos et al., 2014).In this work, EXP is defined as the level of satisfaction of the company with the fulfillment of its economic and strategic objectives related to exports (Cavusgil & Zou, 1994).Based on this definition, as well as the existing consensus in the field of international business, we consider EXP as a concept that comprises several elements, specifically, the achievement of export objectives in terms of sales, profitability, and market share; the export intensity or the percentage exports represent over total sales; and the level of satisfaction of the company with the general performance of its export activity.Thus, EXP was measured using a scale composed of four items adapted from those used by Zou et al. (1998), Lages et al. (2008) and Ahimbisibwe et al. (2016).
Respondent could answer each item using a 7-point Likert-type scale, with 1 = "not at all" and 7 = "to a maximum extent".

| Explanatory variable: CSR under stakeholder view
The explanatory variable of this study is CSR.Based on prior research (Luo et al., 2017;Öberseder et al., 2014), we consider CSR as a construct with six dimensions related to key stakeholders that comprise: shareholders, employees, customers, farmers, community, and competition.Furthermore, we include the environmental dimension as an additional component of CSR.This is because of the intrinsic relevance that environmental issues have in the agri-food sector and its significant interest from a sustainable development point of view (Elkington, 1998).A total of 34 items adapted from Öberseder et al.G omez-Melero (2015) and Cegarra-Navarro et al. (2016).Respondent could answer each item using a 7-point Likert-type scale, with 1 = "strongly disagree" and 7 = "strongly agree".

| Moderation variable: legal form of organization
The legal form of organization is examined as a moderator in our model.The LFO is defined as the legal identity under which a company chooses to carry out its economic activity, affecting its accounting and tax obligations and its liability to third parties.This work distinguishes between associative (i.e., cooperatives) and nonassociative (i.e., capital companies) legal forms.Thus, cooperative societies have a legal form, which is open to all those willing to use their services and partners actively participate in the definition of policies and in the decision making.This LFO is based on the so-called "cooperative principles" (International Cooperative Alliance, 1995) such as democratic participation, education, and commitment to the social issues.LFO was measured using an item that asked participants to report the type of their company's LFO, that is, associative or nonassociative.

| Statistical technique
The structural equation modeling procedure was used to validate the model.Of the different techniques used by the structural equation models for parameter estimation, the PLS technique was chosen, as it is recommended in studies with complex research models, relatively small but representative samples, and in models in which both reflective and formative indicators are included (Hair et al., 2019), as is our case.SmartPLS software was used to carry out the analysis.

| Identification and evaluation of the measurement model
The model is composed of 43 items that measure a total of nine firstorder variables, of which seven correspond to the dimensions of CSR, which is analyzed as a second-order construct.According to the criteria of different authors (Diamantopoulos & Winklhofer, 2001;Jarvis et al., 2003), the items have a reflective relationship with their corresponding variables, except in the case of items that measure the dimensions of CSR.In this case the relationship is formative since these items refer to different aspects of the concept in each dimension which are not necessarily interchangeable and are not expected to be correlated with each other.To evaluate the measurement model, the Build-Up approach technique was used, suggested by Chin (2000) and Aldás-Manzano (2013) for those models that incorporate second-order constructs.This technique proposes, as a first step, to eliminate the second-order factor (CSR) and to relate all its dimensions with all the other constructs to which this second-order factor would be related.For the variables with reflective indicator (EXP and IN), the measurement model was evaluated by assessing the individual reliability of each item; its internal consistency, through Cronbach's alpha and composite reliability; and convergent validity, through the average extracted variance (AVE), developed by Fornell and Larcker (1981).
The results of the estimation process (see Table A1) allow us to conclude that there is internal consistency and convergent validity in all the factors with reflective indicators of the model (EXP and IN).The loads of the items with their respective latent variable exceed the threshold of 0.707 suggested by Carmines and Zeller (1979) and are all significant (p < 0.001).Cronbach's alpha and the composite reliability of all variables exceed the threshold of 0.80 established by Nunnally (1978).The average extracted variance (AVE) exceeds, in all cases, the threshold of 0.50 established by Fornell and Larcker (1981).Discriminant validity was assessed and supported by applying the Fornell-Larcker criterion (i.e., the comparison between the square roots of all constructs' AVEs and the correlations among all constructs) (Fornell & Larcker, 1981) and the heterotrait-monotrait ratio (HTMT) of the correlations proposed by Henseler et al. (2015) (see Table 1).
Regarding the variables with formative indicators (i.e., CSR dimensions), it was found that there is no collinearity between them by using the tolerance and variance inflation factor (VIF) (Diamantopoulos & Winklhofer, 2001).The weights and loads of each indicator were calculated using the bootstrapping technique and its relative relevance was assessed by applying the criteria established by Hair Jr et al. (2014).
It was found that there is no collinearity between the formative indicators in the model, since the VIF is in all cases below the maximum threshold of 5 (Hair et al., 2011).The results of the application of the criteria by Hair Jr et al. (2014) for checking the relative relevance of these indicators (see Table A1) show that the indicators "Com1", "Com2", "Emp1", "Emp4", "Emp5", "Emp7", "Env1" and "Par1" have non-significant weights and their loads are less than 0.50 and non-significant.According to the recommendations of Hair Jr et al. (2014), we removed these indicators, with the exception of the indicators "Com2" and "Emp5" as their loadings are relatively high (close to 0.50) and provide relevant contribution to the constructs and thus preserve content validity.
The second step of the referred Build-Up approach technique consists in converting the original second-order factor CSR to a first-order factor, so that it is measured by its dimensions that now become indicators.In this way, all the variables in the model now have reflective indicators (see Figure 1).Internal consistency, convergent validity and discriminant validity were checked again through the PLS estimation process with overall satisfactory results.Figure 1, Tables 2 and 3 show these results.

| Mediation of IN on the CSR-EXP relationship
The procedure proposed by Baron and Kenny (1986) was used to examine whether the mediation effect of IN on the relationship between CSR and EXP can be supported.First, the direct effect of CSR on EXP was found to be significant (β = 0.455; t = 5.228) in the absence of IN mediation.It was also found that the independent variable CSR significantly affects the mediating variable IN (β 12 = 0.734; t = 18.210) and this, in turn, is significantly related to the dependent variable EXP (β 23 = 0.478; t = 3.032).Finally, it was confirmed that the independent variable CSR no longer significantly affects the dependent variable EXP (β 13 = 0.085; t = 0.635) by incorporating the mediating variable.It can therefore be concluded that IN mediates the relationship between CSR and EXP.
To assess how much of the total effect of CSR on the EXP is due to the mediation of IN, the 'Variance Accounted For' (VAF) criterion was used.Based on this criterion, it is considered that: (1) there is full mediation when the value of the VAF is above the 80% threshold; (2) there is partial mediation when the VAF is between 20% and 80%; and (3) no mediation exists when the value is below the 20% threshold (Hair Jr et al., 2014).In this case, the VAF value was 80.5% indicating that there is complete mediation of IN in the structural model.
Hypothesis 4 is therefore confirmed.

| LFO moderation: a multigroup analysis
The multi-group analysis allows to verify if there are significant differences in the value of β of the CSR-EXP and CSR-CI relationships, according to their LFO.Prior to this analysis, the procedure to assess measurement invariance of composite models (MICOM) was developed (Henseler et al., 2016), which required the following elements to be examined: (1) configural invariance, (2) compositional invariance, and (3) the equality of composite mean values and variances.Configural invariance was automatically established by running MICOM in SmartPLS.Regarding compositional invariance, the p-values of the permutation correlations are greater than 0.05, so the composite invariance is established for all the variables.Furthermore, there are no significant differences in the means and variances of the latent variables in the two groups (p > 0.05).Therefore, a complete invariance of the measurement instrument is supported.
Table 4 shows the results of the multi-group analysis using two non-parametric methods considered as the most reliable to examine the differences in path coefficients between groups (Hair Jr et al., 2017): the permutation method and the PLS-MGA method.
According to the permutation method, there are significant differences if the p-value of the differences between the path coefficients is lower than 0.05.In the PLS-MGA method there are significant differences between the path coefficients of the two groups when the p value is lower than 0.05 or higher than 0.95 (Hair Jr et al., 2017).
T A B L E 2 Internal consistency and convergent validity.As can be seen, there is no significant relationship between CSR and EXP in any of the groups of companies.Moreover, the difference between the two coefficients is not statistically significant either (p > 0.05), so Hypothesis 5a cannot be accepted.On the contrary, the standardized coefficients of the relationship between CSR and IN for the two groups are significant and the p value of the difference between both coefficients is less than 0.05.This shows that the positive impact of CSR on IN is significantly higher in cooperatives than in capital companies.Therefore, Hypothesis 5b is confirmed.

| Theoretical implications
The results of this study offer significant contributions to the literature on the relationship between CSR and EXP in at least five ways.
First, the results show the existence of a close link between CSR and IN, which is in line with the conclusions of previous works (e.g., McWilliams & Siegel, 2000;Surroca et al., 2010).Our results show the existence of a complete mediation of IN in the relationship between CSR and EXP, which implies that almost the entire effect of CSR on EXP is due to IN.This finding goes a step further in the contributions of previous studies that have demonstrated a positive relationship between CSR and EXP (Antonietti & Marzucchi, 2014;Boehe & Barin Cruz, 2010;Xu et al., 2018), since, even though our results demonstrate that the direct relationship between CSR and EXP is not significant, it is shown that it is through IN that companies can incorporate CSR-oriented attributes into their products, their processes and their organizational systems, which in turn helps to increase their EXP.Academics have recognized the potential of CSR as a lever for competitive differentiation and EXP (Barin Cruz et al., 2015;Boehe & Barin Cruz, 2010;Motlaghi & Mostafavi, 2017).(Wood, 1991).
Second, another important contribution of our results is that the mediation effect of IN on the relationship CSR-EXP is significantly stronger in the case of companies with associative LFOs.Prior studies, such as Marcuello and Saz (2008), have stated that the cooperative principle of education, training and information favors the capacity to innovate.In a similar vein, Bakaikoa et al. (2004) argued that cooperation encourages the creation of business networks or industrial clusters, which favors IN.Our findings support the notion that CSR and IN are two dimensions, in cooperative legal forms, which can stimulate the development of intangible assets and, thus, the achievement of greater profitability (Castilla-Polo et al., 2018;Forcadell, 2005).The results of this study add empirical evidence to the literature confirming the moderating effect of LFOs on the relationship between CSR and IN, and contributing to calls for empirical research into the internal mechanisms that influence the relationship between CSR and performance (Tang et al., 2012).At this point, our results confirmed that associative LFO could be an internal factor that strengthens the relationship between CSR and IN.
Third, concerning our contribution in prior literature drawing upon the international business contexts, sometimes it is argued that the existing capabilities in a company are not enough to translate the application of CSR principles into a value proposition for international markets (Costa et al., 2015), which is relevant to sustain or build a competitive advantage abroad.In fact, there is an open question in academic literature that asks why some CSR practices have had a positive and significant impact on EXP improvement, whilst other management practices have not (Boehe & Barin Cruz, 2010).At this point, our results supported the study of Boehe and Barin Cruz (2010) who answered for this question arguing that, in the international competitive context, other management practices may be easier to imitate than CSR practices.In this regard, we show that to be competitive and achieve greater EXP, IN represents a critical factor to apply a CSR orientation in the case of exporting firms that operate in the agri-food sector.This finding provides new evidence in elucidating the missing links that can explain the relationship between CSR and firm performance (Bocquet et al., 2017) and, specifically, the CSR-EXP relationship in the international agri-food sector, which has been scarced analyzed in prior research, despite its significant importance.
Fourth, we also find that IN is a key antecedent for EXP, which is in line with the results of previous studies (Aaby & Slater, 1989;Azar & Ciabuschi, 2017;Damanpour et al., 2009;Lefebvre et al., 1998).

Although companies tend to develop technological and organizational
INs simultaneously, the literature on the relationship between IN and performance has focused mainly on technological INs (Damanpour & Aravind, 2011), which tend to be more visible and less complex to implement.However, recent studies (Azar & Ciabuschi, 2017;Donbesuur et al., 2020) This conceptualization of CSR, fueled by Stakeholder Theory, has been linked to EXP, a measure of performance especially appropriate in the case of export companies.

| Managerial implications
The findings of this work have also significant contribution for practitioners.First, our result support that CSR practices in isolation are not directly connected with improvements in EXP, but the mediation of IN is required.Managers should notice that this combination of organizational variables is a more difficult way for competitors to imitate than the monolithic adoption of a CSR practice.Specifically, the general adoption of CSR strategies without specific transmission channels directly involved with firm performance, which possess greater differentiating potential, entails greater transversality that materializes in companies in an uneven and heterogeneous way depending on their organizational structure and its management model.
Second, CSR should be considered as a strategic and permanent commitment of companies, which is established not only as a response to market pressures, but also as an internalized response (Aguinis et al., 2020).In the situation of a global pandemic, the internalization of CSR practices can help managers to face the challenge of the harmful economic consequences related to COVID-19 (Aguinis et al., 2020).The results show that, in times of high uncertainty in the business sphere, such as the current COVID-19 pandemic situation, managers should not give up CSR practices if they intend to maintain or improve their EXP.
Finally, our findings suggest that it is essential, for CSR-oriented export firms, to develop internal capabilities that foster IN (e.g., adjusting its internal organization or generating and implementing new practices, processes, or techniques) as a key driver to enhance EXP.We recommend to managers of exporting companies that they should invest in the development of IN in socially and environmentally responsible products that will afford them differentiation advantages in foreign markets.Parallel to this strategy of differentiation at the product level, it is necessary for managers to develop a differentiation strategy at the company level so that it is recognized as being socially and environmentally responsible by its customers and other stakeholders in international markets.This, in turn, requires investment in a communication strategy, both internally and externally, regarding the company's commitment to CSR beyond market requirements, focusing on environmental issues (e.g., energy efficiency, efficient use of water, fertilizers and phytosanitary products, waste treatment) and social issues (e.g., the company's contribution to the job satisfaction of its employees, customers and farmers).Such investment is particularly important in contexts where the target markets have high environmental standards such as the international agrifood industry.In fact, companies should emphasize meeting the expectations of both their local and international stakeholders.An international market orientation forces companies to be more competitive because they have to meet a more varied and rigorous demand, especially in terms of CSR (e.g., consumers in Northern Europe demand more ecological and socially responsible products).

| LIMITATIONS AND FURTHER RESEARCH
This study is not exempt from certain limitations which, at the same time, may be the basis for future research.Firstly, the agri-food sector in south-eastern Spain has been used as a research context and its specific characteristics make it particularly suitable for the study of model relationships.Although this may help to make the results more consistent, it also limits their generalization to environments with similar characteristics.Future studies will be responsible for validating the results of this work in other sectors and geographical locations with different features from those proposed in this paper.Secondly, the model has not considered the potential mediating effect of other intangible assets of the company, besides IN, such as reputation, culture, human capital, etc. (Motlaghi & Mostafavi, 2017;Surroca et al., 2010).Future work should examine the mediating or moderating role of these variables in the relationships proposed in this study.
Thirdly, the hypotheses have been tested with data based on the perception of managers that may be conditioned by psychological biases.
Although several actions were taken to control the biases, the fact is that their elimination is sometimes virtually impossible, so the results should be considered with some caution.One opportunity for future research, arising from this limitation, could be to test the hypotheses of this study through the perceptions of other stakeholders.Finally, the data in this paper are of a cross-sectional nature, which prevents the recording of some of the effects derived from the application of CSR practices that differ over time.It would be interesting to carry out research with longitudinal data and cross-sectional delay analysis to exclude potential reverse causality bias and to study the relationships between variables over time.

APPENDIX A
T A B L E A 1 : Descriptive statistics and evaluation of the measurement model.
. In the export scope, IN can provide companies with competitive advantages that help them to increase the volume and profitability of their exports (Pla-Barber & Alegre, 2007).In fact, some studies that have analyzed the IN-EXP relationship under the lens of Stakeholder Theory, found that through investment in IN in products and processes, EXP can be increased because a greater number of stakeholder demands are satisfied (Garcés-Ayerbe et al., 2019), especially of foreign stakeholders argued that the relationship between CSR and performance is indirect and based on the company's intangible resources, including IN among them.Similarly, Bocquet et al. (2017) conclude that companies that adopt CSR practices achieve growth through product and process IN.Reverte et al. (2016) find that IN mediates the relationship between CSR and the performance that includes both financial and non-financial variables.These results suggest that companies that adopt CSR practices achieve better results through their IN, which they use in an instrumental way to obtain positive results.In line with this argument, meaning the positive link between CSR and IN, as well as the positive effect of IN in EXP, we propose that IN is an intangible resource which acts as a mediating variable in the relationship between CSR and EXP.H4.: IN mediates the relationship between CSR and companies' EXP.
) argued that cooperation could be seen as a mechanism that allows stakeholders engagement through evaluation, learning, and sharing experiences.In fact, in the case study of Mondragon Corporation, one of the biggest cooperatives worldwide,Forcadell (2005) argued that IN is one of the principal enablers of the management model of cooperativism since IN provides support for customers and facilitates focusing resources on the right way.Thus, because of the combination of strong democratic values (i.e., "the cooperative principles") related to CSR, joined to the important role played by IN on profitability for cooperatives, we consider that there is a moderating effect of the LFO on CSR-IN relationship.This relationship will be more intense in the case of cooperative legal forms as their commitment to CSR will be more consistent.Thus, we hypothesized:H5b.: The LFO moderates the relationship between CSR and IN, making it more intense in the case of cooperative legal forms.

(
2014),Dopico et al. (2012),Luo et al. (2017),Turker (2009), andFatma et al. (2014) were used to create and measure this multidimensional construct of CSR.Respondent could answer each item using a 7-point Likert-type scale, with 1 = "strongly disagree" and 7 = "strongly agree".4.2.3 | Mediation variable: ININ acted as mediator in the proposed model.At this point, we adopted a broad conception of IN attempting to capture the four domains ofIN (i.e., product, process, marketing, organizational)  that are proposed in the Oslo Manual(OECD, 2005).In this sense, both technological and organizational INs are necessary to better understand the way in which companies face the changes and uncertainties of their environment(Damanpour & Aravind, 2011;Polder et al., 2010).Technological INs are directly related to the company's main activity and its operating systems (e.g., development of new goods and services or the improvement of existing ones, adoption of new or improved technological production and distribution methods), while organizational INs mainly affect its management systems (e.g., development of new management techniques, new organizational structures, new strategic orientations, INs in markets or marketing methods) (OECD, 2005).The two types of IN complement each other, so that the generation of technological IN is conditioned by the company's capacity to integrate new technology (Lam, 2004).Thus, organizational INs create an appropriate environment and facilitate the adoption and efficient use of technological INs.In this work IN was measured using a scale composed of five items referred to technological and organizational INs, extracted and adapted from Gallardo-Vázquez and Sánchez-Hernández (2013), The values on the diagonal (in italics) are the square roots of the average extracted variance; below the diagonal are the correlations between factors; above the diagonal is the heterotrait-monotrait ratio.F I G U R E 1 Structural diagram after application of the Build-Up approach technique.etal., 2016).The results indicate that there is no relationship between CSR and EXP (β = 0.085; f 2 = 0.005; t = 0.635), and therefore Hypothesis 1 cannot be accepted.However, the effect of CSR on IN is positive, high, and statistically significant (β = 0.734; f 2 = 1.167; t = 18.21), so Hypothesis 2 is confirmed.IN has a positive and significant impact (β = 0.478; f 2 = 0.150; t = 3.032) on EXP, thus confirming Hypothesis 3. Figure1show these results.The values of R 2 , Q 2 , and predictive Q 2 indicate that the model has predictive capacity or , argue the importance of considering technological and organizational INs together when studying the IN-EXP relationship.Both INs need to be aligned to optimize the effect of the IN on the EXP (Donbesuur et al., 2020).This study adopts a broad conceptualization of the IN that includes both technological and organizational IN indicators, rather than focusing on a single category, demonstrating the predictive power of these aspects of IN in EXP under the effect of CSR.Finally, prior studies drawing on Stakeholder Theory have stated that the main advantage of CSR to create firm value resides in its capability to generate positive stakeholders' relationships for the firm(Laplume et al., 2008;Peloza & Shang, 2011).From a conceptual point of view, this work contributes to the previous literature on Stakeholder Theory by providing a broad and multidimensional conceptualization of CSR that includes internal dimensions (employees and shareholders), external dimensions (customers, farmers, community, and competition), as well as the environmental dimension due to its special relevance in the context of this research.

(
company has introduced innovations in marketing methods in the last two years 0.886** (In5) Our company has introduced new practices that enhance the capture of new international markets 0.825** (i) Export performance (Exp1) Over the past few years, out export sales have increased 5.292 (1.303) N/A N/A 0.928** CA = 0.942 CR = 0.959 The values on the diagonal (in italics) are the square roots of the average extracted variance; below the diagonal are the correlations between factors; above the diagonal is the heterotrait-monotrait ratio.Multi-group analysis results.
Our company implements flexible policies to provide a good balance between work and life for its employees Far5) Our company promotes organic and/or integrated production among its farmers/producers (Far6) Our company considers it essential that its farmers/ producers have certification of good agricultural practices, food safety certification and environmental certification (Continues)T A B L E A 1 (Continued) (