2.1 Innovation and Sources of Knowledge

According to the resource-based view (RBV) of a firm (Wernerfelt, 1984), the resources and capabilities available in a company determine its ability to innovate, and knowledge is the most fundamental resource to achieve innovation (Farooq, 2018). With the open innovation paradigm (Chesbrough & Bogers, 2014) the company can increase its knowledge for innovation through the knowledge flows it can receive from other companies and organizations.

Sources of knowledge are the flows of knowledge that come from actors who share resources and create value for innovation in companies, such as customers, suppliers, competitors, consultants, universities and other R&D institutions that are part of a national innovation system (Freeman, 2004; Lundvall, 1992).

Internal sources comprise the internal knowledge that the company has for innovation: they collect the knowledge acquired in R&D and the knowledge available to the company's departments, being the most important knowledge for innovation (Gu et al., 2016; Ruiz-Pava & Forero-Pineda, 2018). When a company is part of a business group, it can receive flows of knowledge and technology from other companies in the business group, and these flows increase the internal knowledge of the company that the company has available for innovation (Kim & Lui, 2015).

Business groups play a particularly important role in the R&D of their subsidiaries because they provide knowledge to adapt new products to the local market, especially for products that are in other markets, or because they contribute to the creation and acquisition of globally relevant technology for the group (Zhang et al., 2015).

Customers contribute to innovation with market needs and experience of the use of products, and this knowledge is the input for the development of new products. It is the most important knowledge to develop innovations because it guides the company to innovation based on what the market needs (Cui & Wu, 2018; Melander, 2020).

Suppliers contribute with technologies in the supply chain that are used in product parts (Kumar et al., 2016), including embedded technology such as machinery and equipment, and non-embedded technology such as patents and licences (Dosi et al., 2021; Moreira et al., 2020).

Competitors provide market knowledge, product technology (Ozkaya et al., 2015) and critical information about the market, technologies and patents (Wang & Gao, 2021). Consultants provide ideas to increase innovation capacity (Doloreux & Turkina, 2017) and help companies receive technology transfers (Bianchi et al., 2016).

Universities contribute to the production of scientific knowledge and technological development, which is especially useful for large companies with high technological knowledge (Bellucci & Pennacchio, 2016; Neves et al., 2021), but the company needs to have internal absorptive capacity to receive this transfer of knowledge and technology (Østergaard & Drejer, 2022).

Public research institutes are the government's support for business innovation. They provide specialized knowledge for specific clusters or sectors of companies, but companies must also have internal absorptive capacity to receive this transfer of knowledge (Cheah & Ho, 2020; Jung et al., 2021).

Internal sources are the most important and are positively related to the company's innovative performance. The knowledge flows for innovation obtained from external sources such as customers, suppliers, competitors, universities and public research institutes increase the company's innovative potential (Laursen & Salter, 2006).

Several studies have examined the individual effect of each source on product innovation: internal sources, customers, competitors and university (Gómez et al., 2016); customer and supplier sources (Moon et al., 2018); consultants who help transfer knowledge from universities and public research institutes (Bianchi et al., 2016); universities that impact product innovation mainly in radical innovations and with large companies (Neves et al., 2021; Pittayasophon & Intarakumnerd, 2017); and public research institutes that develop knowledge for business sectors and impact product innovation (Cheah & Ho, 2020).

In their study in Peru, Moya-Fernández and Seclen-Luna (2024) studied the impact of each source of knowledge and found that internal sources, customers and consultants are positively related to product innovation, while Arenas and Gonzalez (2019) found that consultants are related to product innovation in Peru.

Each source of knowledge has an impact of different intensity depending on the characteristics of the industrial sector in which the company operates and the competitiveness of each country (Gómez et al., 2016; Pejić et al., 2015). Likewise, the impact of each source of knowledge has a different intensity depending on the technological level of the sector and the complexity of the knowledge managed by the company (Doloreux & Turkina, 2017; Gu et al., 2016).

2.2 Knowledge Source Combinations for Innovation

Knowledge sources are heterogeneous-that is, each provides companies with a different type of knowledge. The company needs to access different sources, and makes the selection and combination based on the type of innovation it wants to carry out and the internal knowledge it has; it needs to select external knowledge that is complementary and can be integrated with the company’s internal knowledge (Chen et al., 2016; Criscuolo et al., 2018; Rodríguez et al., 2016).

The firm has knowledge components, each of which is associated with a technological or scientific concept, and the recombination of the various knowledge components gives rise to new meanings or functions that are the basis for achieving innovation (Xiao et al., 2022). Knowledge recombination in the firm can generate new knowledge or the reuse of existing knowledge, leading to innovation (Appio et al., 2017).

The firm prioritizes its open innovation strategy, selecting the different sources it can access to obtain knowledge based on learning from past collaborations, its absorptive capacity, and the heterogeneity of the external knowledge it can access (Kesidou et al., 2022). Knowledge combination is more advantageous in firms in high-tech sectors (Wu & Wang, 2017). A firm's knowledge-creating capacity is positively associated with its absorptive capacity (Khraishi et al., 2022).

Obtaining knowledge for innovation, which comes from the integration of internal knowledge with external knowledge obtained from different sources, is achieved through joint action of the company's internal capabilities, especially R&D capacity and the knowledge available for innovation (Criscuolo et al., 2018; Ma & Yu, 2021).

Table 1 shows studies that have examined the relationship between combinations of knowledge sources and different types of innovation. A first research trend was to examine the impact of external sources of knowledge (customers, suppliers, competitors, universities, public research institutes, etc.) on different types of innovation such as product and process innovation, and service innovation. The evidence indicates that using different knowledge sources creates a combination effect that is distinct from the effect of each knowledge source separately (Rodríguez et al., 2017).

The studies carried out showed various combinations-customers and competitors, or customers and suppliers, or universities and research institutes-can increase the innovative potential of the company, and this external knowledge is complementary to the internal knowledge of the company (e.g. Chen et al., 2011, 2016; Cruz-González et al., 2014; Laursen & Salter, 2004; Leiponen, 2005; Tödtling & Grillitsch, 2015; Rodríguez et al., 2016).

A second research trend was to examine combinations of internal sources with external sources of knowledge where internal sources have an important role and the company's internal capabilities. These combinations are: internal with suppliers, customers and competitors; internal with suppliers, competitors and universities; internal with customers, competitors and universities (e.g. Criscuolo et al., 2018).

The impact of external sources was also examined according to the degree of innovation novelty (new-to-firm and new-to-market product innovations), Köhler et al. (2012) found differences between the combinations appropriate for new-to-market product innovations and those appropriate for new-to-firm product innovations; the combination of suppliers with professional associations and the combination of universities and public research institutes are more appropriate for new-to-market product innovations, and the combinations of customers and competitors are more appropriate for new-to-firm product innovations.

Sarpong and Teirlinck (2018) found that collaboration with customers or competitors, is appropriate for new-to-market and new-to-firm product innovations, and collaboration with universities or consultants or private/public research institutes and collaboration with suppliers are appropriate for new-to-market product innovations.

Other combinations that have been found appropriate for new-to-market product innovations are: combination of customer and suppliers, and combination of universities and research institutes (Duodu & Rowlinson, 2021) combination of internal sources with customers (Duran et al., 2022), combination of customers, competitors, suppliers, and firms from different industries; combination of universities and research centres (Ruiz-Pava & Forero-Pineda, 2018).

Most of the investigations in Table 1 focus on product and process innovation, and in developed countries. Differences are shown between the combinations of one country and another country, and between industrial companies and service companies. Evidence of combinations of knowledge sources in emerging countries is limited.

Product innovation includes both incremental innovation and radical innovation, but radical innovation is more elaborate; it requires more specialized knowledge than incremental innovation, which is to improve products, and requires a lower level of knowledge in the company (Gui et al., 2022).

New-to-market product innovation (national or international market) includes both radical new product innovations and new products for the national market that are imitations of other products that exist in other markets. It is more complex than new-to-firm product innovation (Ritala et al., 2018). New-to-market product innovation is more elaborate; it requires greater efforts to integrate internal and external knowledge to achieve innovation (Doloreux et al., 2018; Jugend et al., 2018), and to be successful in combining knowledge sources requires strong capabilities and cooperation with internal sources (Hervas-Oliver et al., 2022; Plechero & Grillitsch, 2023; Rodríguez et al., 2017).

The industries of Peru, using the classification of manufacturing industries based on technology (OECD, 2007), correspond mainly to low-technology industries and medium-low-technology industries. They are different from companies in developed countries, where there is generally evidence of the aspects that influence the combinations in terms of the level of complexity of the knowledge that the industry handles and the complexity of the innovations (Kesidou et al., 2022; Wu & Wang, 2017).

The conclusions of combinations found in developed countries can only be generalized to high-tech manufacturing firms from developed countries, since they have been obtained in other contexts of technology companies, other industrial sectors and with greater knowledge complexity (Cruz-González et al., 2014).

There is little knowledge about the combinations of knowledge sources that influence the degree of novelty or radicalness of innovations (Criscuolo et al., 2018). It is necessary to expand knowledge of how companies combine the knowledge components they have to achieve innovation (Xiao et al., 2022).

Therefore, there is a gap in the literature about the combinations that influence the degree of novelty or radicalness of innovations (new-to-firm and new-to-market product innovation), especially in emerging countries. In order to contribute to this gap in the literature, this research raises the following hypotheses.

Considering that companies can combine sources that include knowledge from internal sources (within the company or business group) that provide ideas and internal R&D knowledge, knowledge from companies in their business group (De Beule & Van Beveren, 2019), knowledge provided by customers about product needs (Cui & Wu, 2018; Melander, 2020), and knowledge provided by suppliers about technologies (Kumar et al., 2016), and that previous studies have found a positive relationship in manufacturing between the combination of internal sources, suppliers and customers and product innovation (Criscuolo et al., 2018), the following hypotheses are argued.

H1a: The combination of internal sources of knowledge (within the company or business group), customers and suppliers is positively related to new-to-firm product innovation.

H1b: The combination of internal sources of knowledge (within the company or business group), customers and suppliers is positively related to new-to-market product innovation.

Considering that firms' innovations may require a higher level of knowledge complexity, which requires knowledge from a greater number of sources, knowledge that suppliers provide about technologies (Kumar et al., 2016) and knowledge of competitors about the market and product technology (Ozkaya et al., 2015; Wang & Gao, 2021) and that previous studies have found a positive relationship in manufacturing between the combination of sources customers, suppliers and competitors, and product innovation (Criscuolo et al., 2018; Del Carpio Gallegos & Seclen-Luna, 2022), the following hypotheses are argued.

H2a: The combination of internal sources of knowledge (within the firm or business group), customers, suppliers and competitors is positively related to new-to-firm product innovation.

H2b: The combination of internal sources of knowledge (within the company or business group), customers, suppliers and competitors is positively related to new-to-market product innovation.

Considering that previous studies in Peru have found a positive relationship in manufacturing companies between consultants and product innovation (Moya-Fernández & Seclen-Luna, 2024) and between the combination of universities and research centres with product innovation (Del Carpio Gallegos & Seclen-Luna, 2022), and that previous studies have identified combinations of universities, research institutes and consultants/R&D laboratories with product innovation (Cruz-González et al.,2014; Haus-Reve et al., 2019), the following hypotheses are argued.

H3a: The combination of knowledge sources from universities, public research institutes and consultants/private R&D institutes is positively related to new-to-firm product innovation.

H3b: The combination of knowledge sources from universities, public research institutes and consultants/private R&D institutes is positively related to new-to-market product innovation.

EFA was carried out to group variables into factors. The KMO index of sampling adequacy had a value of 0.715, values greater than 0.60 considered appropriate (Kim & Mueller, 1978), and the Bartlett's sphericity test had a significance level of 0.000, being less than 0.05 (Kim & Mueller, 1978). The extraction method used was the principal components rotation method: varimax with Kaiser normalization. Factors with an eigenvalue greater than one were extracted (Fabrigar et al., 1999; Hair et al., 2010). The communalities obtained were >0.60, and the factor loading of each variable was greater than 0.5, which are appropriate values (Fabrigar et al., 1999). In addition, each factor is composed of at least three variables (Fabrigar et al., 1999). The variables that make up each factor can be seen in Table 4. The second column shows the number of firms using the variable.

Two factors were obtained from the factorial analysis: the first was made up of internal sources (within the firm or business group), customers, suppliers and competitors, called internal and market sources; the second factor was made up of university sources, research institutes and consultants/private R&D institutes, called science source and consultants.

CFA was carried out to verify that the factors obtained in the EFA have an adequate level of confidence. It was verified that the factor loadings were greater than 0.50. In addition, the average variance extracted (AVE) was obtained: it was greater than 0.50. For the construct internal and market sources, Cronbach's alpha of 0.71 was obtained, and for the construct science source and consultants, Cronbach's alpha of 0.75 was obtained, which are considered appropriate measures of reliability.

Absolute fit indices are applied that indicate the degree to which the model observed in the covariance matrix matches the covariance matrix of the implicit model. Values of RMSEA = 0.056 and SRMR = 0.04 were obtained, ​​indicating an adequate fit of the model.

In this way, CFA confirms the relationship between the internal and market sources factor and the observed variables-internal sources (within the firm or business group), customers, suppliers and competitors-and the science source and consultants factor and the observed variables-university, research institutes, and consultants/private R&D institute, which were found in the EFA

The model was examined for endogeneity, multicollinearity and heteroscedasticity problems. A strong test for non-multicollinearity was used, where VIF values ​​less than 3 were obtained. Pearson correlations between the variables were examined as shown in Table 5: as can be observed, the correlations are very low, less than 0.5. Table 6 shows the results for the bivariate probit regression, where robust errors were used.

The objective of this research was to examine the combinations of knowledge sources that influence the degree of novelty or radicalness of innovation in Peru. In line with this objective, the relationship between the sources of knowledge used by the company and new-to-market and new-to-firm product innovation is examined.

The results show two combinations. Firstly, the combination of internal sources (within the company or business group) with the knowledge of customers, suppliers and competitors is related to new-to-firm and new-to-market product innovation. Secondly, the combination of knowledge from universities, public research institutes and consultants/private R&D institutes is related to new-to-market product innovation.

In the first combination, the internal knowledge source (within the company or business group) can contribute with knowledge, know-how and even R&D results from other companies in the group (Hsu et al., 2015; Zhang et al., 2015). Of the companies examined in Peru, 22.3 per cent are part of a business group. Competitors are important as they have knowledge of the market and product technologies, while customers contribute with their needs and their experience with the products. This is complemented by the knowledge of competitors in the market, which is very important, especially for innovations that adapt products for the local market from products that already exist in other markets. Suppliers contribute with knowledge of part technologies.

Regarding the first combination, the results are found to be different from the combinations generally found in developed countries for product innovations and processes innovation, such: internal sources with customers and suppliers (e.g. Chen et al., 2016; Criscuolo et al., 2018; Köhler et al., 2012), internal sources with competitors (e.g. Criscuolo et al., 2018; Rodríguez et al., 2017).

This first combination is similar to the combination of internal sources, customers, suppliers and competitors for product innovation presented by Criscuolo et al. (2018) in the UK, whose authors found other combinations, such as internal with suppliers, competitors and universities; internal with customers, competitors and universities.

Regarding the second combination, the results found are shown to be different from the combinations of knowledge that include universities and research institutes (e.g. Chen et al., 2016; Ruiz-Pava & Forero-Pineda, 2018), since the combinations found include universities, public research institutes and consultants/private R&D institutes. The consultants help compensate for the internal capabilities of the companies, such as absorptive capacity, while private R&D institutes help develop R&D activities.

This research differs from others developed in emerging countries for product innovation, Del Carpio Gallegos & Seclen-Luna (2022) in Peru found the combination of customers, suppliers and competitors, and the combination of universities and research centres. Escorcia-Caballero et al. (2022) in Colombia found combinations of customers, suppliers, companies from other sectors and associations, and the combination of universities, research and productivity centres, business incubators, technology parks and specialized consultants. However, these combinations focus on examining external sources only and are for product innovation (radical and incremental).

The current study also differs from the research by Ruiz-Pava and Forero-Pineda (2018) in Colombia that found an exclusive combination with internal sources; a second composed of customers, competitors, suppliers and firms from different industries; and a third composed of technological development centres, research centres, universities and training centres. However, the research focuses on studying combinations for innovation by type of market, national or international, which is different from the combinations for new-to-firm and new-to-market product innovation. Also, these combinations do not include consultants.

This research contributes in the following ways. Firstly, it provides insights into the combinations that influence the degree of novelty or radicalness (new-to-firm and new-to-market product innovation) for Peru, an emerging country.

Existing theory on knowledge source combinations indicates that companies make combinations influenced by the type of innovation and the firm's internal capabilities (Criscuolo et al., 2018), new-to-market product innovations includes radical innovations that are more complex than those found for product innovations, in addition the companies examined have particular characteristics, correspond mainly to low-technology industries and medium-low-technology industries and have an average of 0.0016 in R&D intensity (R&D expenditure/sales) which implies that they have lower internal capabilities to generate knowledge through R&D.

The existing combinations generally obtained in developed countries can only be generalized to high-tech manufacturing firms from developed countries (Cruz-González et al., 2014) and the combinations found contributes because knowledge about the combinations that influence the degree of novelty or radicalness (new-to-firm and new-to-market product innovation), especially in emerging countries, is still scarce.

Secondly, this research contributes evidence by including in combinations knowledge from other firms in the business group of which the firm is part. Evidence from emerging countries on combinations has generally considered knowledge from internal sources without considering that the company may receive knowledge from other companies in the business group. The Peruvian survey is aligned with the Oslo Manual (OECD/Eurostat, 2018), so it already records internal sources as knowledge within the firm or business group.

Thirdly, the study provides evidence on the role of consultants/private R&D institutes in combinations in emerging countries. In Peru, consultants/private R&D institutes play an important role in new-to-market product innovations: universities and public research institutes generate knowledge, and the role of consultants/private R&D institutes is to contribute to the implementation of knowledge from universities and public research institutes.

The evidence that exists from developed countries has rarely included consultants/ private R&D institutes in combinations for new-to-market product innovations. This is because universities in developed countries play an important role, especially in developing technologies used in radical innovations in high-tech companies. Universities and public research institutes have greater capabilities for technological development, while companies have greater capabilities and resources for directly implementing this knowledge.

Finally, the research contributes because the combinations found represent the external knowledge search strategies employed by these companies and the sources of knowledge they combine to achieve new-to-firm and new-to-market product innovations, which is of practical use to companies and public policymaking bodies.