Responsible Innovation (RI) is generally defined as the innovation of “taking care of the future through collective stewardship of science and innovation in the present” (Stilgoe, Owen, & Macnaghten, 2013). The research on RI that emerged over the last decade has focused on the rationale for RI and the institutionalization of RI in policy (Flink & Kaldewey, 2018; Genus & Iskandarova, 2018; Owen, Pansera, Macnaghten, & Randles, 2021). Recently, the growing body of literature on organizing RI in business (Jarmai, 2020; Ogbeibu et al., 2021; Scherer & Voegtlin, 2020) has evolved as management scholars explore the potential contributions of businesses to addressing grand challenges, such as climate changes, sociopolitical complexity and uncertainty, poverty, and pandemics (Brammer, Branicki, Linnenluecke, & Smith, 2019; Buckley, Doh, & Benischke, 2017; George, Howard-Grenville, Joshi, & Tihanyi, 2016). Although the literature considers the critical role of RI for sustainability and societal norms about science and innovation in business, the determinants of organizing RI in business are still underexplored and need further research. The theoretical gap appears to be because prior studies have largely concentrated on the conceptual meaning of “moral responsibility” in science and innovation: for example, inclusive responsibility (Cai, Gippel, Zhu, & Singh, 2019; Malhotra, Majchrzak, & Niemiec, 2017; Malhotra & Van Alstyne, 2014) and reflexive responsibility (Singh, Chen, Del Giudice, & El-Kassar, 2019; Singh, Giudice, Chierici, & Graziano, 2020). In other words, previous research on RI has explored how moral responsibility can be enhanced in science and innovation from the stakeholder’s view.

However, while this view has its own merits, merely emphasizing the roles of stakeholders and corporate social responsibility (CSR) may be insufficient to fully understand the characteristics of organizing RI in business to achieve both responsible and innovative management. The paper aims to explore the determinants of organizing RI and contributes to the literature in the Asian business context, extending the relatively skewed theoretical attention to a holistic view that combines both responsibility and innovation management. First, based on the work of Stilgoe et al. (2013) on the four dimensions of the RI concept, such as anticipation, reflexivity, inclusion, and responsiveness, we identify two primary determinants of RI in business as follows: (1) organizational agility to conduct innovation for anticipation and responsiveness and (2) communicative actions to enhance responsibility with stakeholders for inclusion and reflexivity. Second, we examine the RI of manufacturing industries in Asia-Pacific regions, where the institutionalization of integrated policies among science, technology, and innovation has recently been highlighted (UNESCAP, 2016). According to the OECD report (OECD, 2017), the population exposed to pollution levels above WHO guidelines in Asia-Pacific countries (e.g., 100% in China, 99% in Korea, 98% in Japan) is much more serious than those in other areas (e.g., 7% in the U.S., 75% in EU, and 2% in Canada). The United States, Japan, Germany, South Korea, and France are the main OECD countries that contribute around 90% of green inventions in the world. Especially, the empirical investigation of South Korean firms’ RI performance is worthwhile to study the underlying mechanism of organizing RI internally at the firm level because South Korea has institutionalized environment-related policies and regulations for decades (OECD, 2017).

Drawing on the theories of organizational agility and the stakeholder’s view, we develop hypotheses about the effect of the combination of internally agile and externally communicative management on RI performance. We tested and found support for our hypotheses, using a large-scale sample of 13,234 manufacturing firms in South Korea during the period between 1999 and 2019. Methodologically, our empirical model includes a new proxy to measure organizational agility for manufacturing firms and uses a zero-inflated negative binomial regression to analyze the count data of the green patents in a large-scale sample. The advantage of our empirical strategy is to offer a new measurement of the process-related agility of manufacturing firms and to contribute methodologically to RI literature related to the use of patent data to examine RI performance. Consequently, the findings show that the combination of organizational agility and communicative actions can enhance RI performance and support the holistic view of RI.

Our study contributes to recent literature on RI in the Asian business context (e.g., Chatterjee, Chaudhuri, & Vrontis, 2021; Magni, Palladino, Papa, & Cailleba, 2022; Ogbeibu et al., 2021; Pan, Yuan, Song, Li, & Wang, 2022; Scuotto, Garcia-Perez, Kalisz, & Dhir, 2022). The activities of manufacturing firms tend to generate both positive externalities (e.g., innovations to contribute to socio-economic welfare) and negative externalities (e.g., pollution to damage environmental sustainability). The implementation of RI in the manufacturing industries is important for countries in the Asian-Pacific region to sustain socio-economic growth and avoid negative externalities inherently. As the concept of RI extends to deeper and broader aspects of integrated dimensions of RI (e.g., anticipation, reflexivity, inclusion, and responsiveness) than the stakeholder management in CSR (Genus & Iskandarova, 2018; Stilgoe et al., 2013), our study contributes to the extension of empirical research on RI in the holistic view and offers managerial implications for organizing RI in business. In the following sections, we justify hypotheses development, methodology, and empirical analysis in turn. Finally, we will discuss the implications of our findings for academics and practitioners.

Theoretical background

Organizational agility and stakeholders for RI

Grand challenges have been considered a salient context in recent management research since the United Nations (UN) announced a list of 17 sustainable development goals (SDGs) to resolve societal, economic, and environmental problems worldwide in 2015 (Sachs & Sachs, 2021; Van Tulder, Rodrigues, Mirza, & Sexsmith, 2021). As the salient context of grand challenges emerges, recent RI research extends from the themes of innovation orientations to avoid harmful outcomes of innovation activities toward the way of finding solutions to solve the grand challenges in business (Arslan & Tarakci, 2022). The importance of RI in the context of grand challenges is related to the theoretical foundation of organizational adaptability in response to an external shock (Christofi, Pereira, Vrontis, Tarba, & Thrassou, 2021; Sanchez, 1995; Teece, Peteraf, & Leih, 2016; Volberda, 1996). Thus, we consider that the essential determinants of organizing RI in business mainly include specific organizational responses under external shocks, such as agile management and organizational action programs that change the orientations of innovation management.

On the one hand, an organizational agility perspective can provide compelling insights into how to organize RI in response to grand challenges that are relatively enormous and constant problems that humans face. In general, strategic agility is defined as “the ability to remain flexible in facing new developments, to continuously adjust the company’s strategic direction, and to develop innovative ways to create value” (Weber & Tarba, 2014: 5). That is, the concept of agility highlights the continuity of the changing mechanism for strategic adaptation, contributing to explore how firms can remain flexible and adaptable continuously to environmental changes (Shams et al., 2021). A recent stream of management research also highlights the continuously flexible status of agile organizations. For example, learning and innovation research focuses on how to transform agile organizations into rapid learning and repeatable decision-making cycles at scale (Brosseau, Ebrahim, Handscomb, & Thaker, 2019; Rigby, Sutherland, & Noble, 2018) and how to perform agile innovation with rapid, iterative feedback cycles in response to the frequent changes in customer preferences and solution options (Rigby, Sutherland, & Takeuchi, 2016).

On the other hand, a stakeholder perspective has emphasized the important role of stakeholders in providing resources to generate potential economic profits, besides the shareholder’s view of profit appropriation in value creation (Barney, 2018; Freeman, Wicks, & Parmar, 2004; Zingales, 2000). Prior studies show that non-shareholding stakeholders, such as suppliers, non-equity alliance partners, customers, and the other public institutions, can allow focal firms to exploit redundant resources and capabilities that might contribute to open innovation (Baldwin & von Hippel, 2011; Chesbrough, 2003) and co-specialized value creation within business ecosystems (Moore, 1996; Rong, Wu, Shi, & Guo, 2015; Teece, 1986). In particular, the stakeholder view has contributed to the literature on RI in terms of corporate social responsibility, focusing on business ethics (Pandza & Ellwood, 2013; Voegtlin & Scherer, 2017) and corporate governance (Scherer & Voegtlin, 2020). For example, when stakeholders are interested in resolving societal and environmental for sustainable development, firms tend to include them in the corporate decision-making process and creating shared values (Porter & Kramer, 2011; Scherer & Voegtlin, 2020). As such, based on corporate social responsibility, RI highlights the normative premise that firms should avoid harm and do good in response to grand challenges (Genus & Iskandarova, 2018; Genus & Stirling, 2018; Stilgoe et al., 2013).

This theoretical foundation justifies that agile management and stakeholder-based management might be considered two primary organizational responses in the context of grand challenges. We build on this foundational idea and advance a theory of RI, examining how these organizational responses can affect RI performance. Especially, we concentrate on the underlying mechanisms of RI performance implications based on the conceptual dimensions of anticipation, responsiveness, inclusion, and reflexivity (Stilgoe et al., 2013) in terms of a holistic view. We build the argument in two directions: an inside-out aspect and an outside-in aspect. First, taking the inside-out approach, we argue that organizational agility allows firms to adjust innovative activities rapidly to grand challenges by proactively setting technological orientation and reallocating resources quickly so as to enhance anticipation and responsiveness for RI. Second, from the outside-in approach, we propose that communicative actions enable firms to involve external stakeholders in internal decision-making processes and to develop social and ethical norms for creating shared values so as to improve inclusion and reflexivity for RI. In addition, we also examine the interaction effect of the two mechanisms on the performance implications. Figure 1 represents our theoretical framework of underlying mechanisms in detail and will provide overall guidelines for hypothesis development.

Fig. 1
figure 1

A Theoretical Model of Responsible Innovation. (Note: Solid lines highlight the relationship between primary dependent and independent variables. Dashed lines indicate the theoretical explanation to address the rationale of RI performance that can be unobservable in a black box.)

Full size image

Hypothesis development

Agile organizations for anticipation and responsiveness

Building upon a strategic flexibility and agility perspective, we predict that agile organizations are more likely than traditional hierarchical organizations to provide the competencies that contribute to the firms’ RI. This effect results from the internal organizational attributes of agile organizations. While traditional organizations govern decision-making processes through a static, structural hierarchy in the context of relatively predictable changes in business environments, agile organizations are rather democratized to grant decision-making rights to persons or teams that are adequate to solve the problems in response to uncertain environments (Brosseau et al., 2019). The organizational design of agile organizations is generally characterized as a continuously flexible status by simultaneously pursuing some paradoxical attributes, such as velocity and adaptability versus stability and efficiency (Aghina, De Smet, & Weerda, 2015; Lewis, Andriopoulos, & Smith, 2014; Weber & Tarba, 2014). Organizational agility can be considered an essential organizational factor for firms to adapt to uncertain environments frequently or continually, which means that organizational agility enables firms to be resilient and rapidly change organizational and strategic directions.

Through the inside-out approach, we argue that this characteristic of agile organizations is associated with the competencies to adjust to the anticipatory and responsive processes for RI in response to grand challenges. In RI, anticipation is not just about exactly predicting or forecasting the future with a passive attitude; rather, it is to shape desirable technological futures and organize resources and capabilities to pursue the goals with a relatively proactive attitude (Stilgoe et al., 2013). For example, manufacturing firms cannot exactly predict what their new manufacturing activities, good or bad, will generate in advance. Instead, they should be able to shape a good orientation to develop and use new technologies that can generate positive externalities. As such, through backward inductionFootnote 1 (i.e., first by setting a good technological objective and then doing in turn what firms need to prepare for it), firms can achieve RI by proactively making brilliant futures of technological development and innovation for human society and the environment. For this reason, the best strategy for firms seeking RI would be to set the right technological futures in advance and quickly organize the necessary resources and capabilities in a timely manner.

Moreover, the implementation of this anticipatory decision-making process and operations depends on how widely and quickly firms can change systems or directions in response to changes in uncertain circumstances. For RI, firms must have the ability to rapidly change the direction of shaping technological futures and continually adjust their activities with the best possible response to lead to co-evolution between technology and society (Genus & Stirling, 2018; Stilgoe et al., 2013). Compared to hierarchical top-down innovation from initial plans to subsequent operations, agile innovation can be conducted by quickly reallocating relevant resources based on incremental thinking, so-called the "test-and-learn approach" (Rigby et al., 2018). In other words, agile innovation activities include using current technologies as much as possible at first and then adjusting prototypes based on feedback from current utilization until new technologies emerge. For example, tech companies like Apple, Amazon, and Tesla have continued to receive customer feedback on new products and services whenever they grow, and their agile innovation processes are iterated to stay good or fix flaws. Consequently, organizational agility can improve the firm’s responsiveness and thereby reinforce RI in response to grand challenges. Based on the reasoning so far, we argue that firms with a high degree of organizational agility are more likely to implement anticipatory processes to shape technological futures in the most responsive way in the context of uncertainties. In light of this, we propose the following hypothesis:

  • Hypothesis 1 The degree of organizational agility will be positively related to the performance of RI.

Communicative actions with stakeholders for inclusion and reflexivity

Based on the perspectives of corporate social responsibility and corporate governance, we suggest that proactive communicative actions will be likely to contribute to RI. This outside-in approach assumes Hayek’s knowledge economy that firms cannot know everything about how to benefit social and environmental welfare due to the very limited knowledge dispersed across all relevant agents in society (Hayek, 1945). As such, we emphasize the broad deliberative processes of the firm in which all relevant stakeholders can participate to extend the knowledge base beyond the boundaries of the firm. The mechanism behind this effect can be conducted in two ways, including how broadly firms can build inclusive decision-making systems (i.e., inclusion) and how deeply such systems can address ethical issues to improve social and environmental welfare around their business (i.e., reflexivity).

First, proactive communicative actions with stakeholders will facilitate openness in decision-making and innovation processes, and thereby firms can meet social needs as closely as possible. Stakeholders can be from a variety of areas, including academia, government, civil communities, and business. Since they may have different interests in social, economic, and environmental problems, the inclusive decision-making process through communicative actions can allow firms to access diverse voices on social issues from different areas related to their business (Scherer & Voegtlin, 2020). Also, broad stakeholder engagement through inclusive corporate governance enhances distributed and collective intelligence for open innovation (Baldwin & von Hippel, 2011; Chesbrough, 2003). Therefore, the results of the innovation will satisfy social needs and values.

Second, continuous communicative actions with stakeholders are helpful for firms to build business ethics closely related to sustainable development. Regarding sustainable development for the profit, the people, and the planet, the so-called “triple bottom line” (Norman & MacDonald, 2004), firms’ proactive communicative actions are critical to encouraging the engagement of stakeholders in the governance of science, technology, and innovation so as to enhance social and economic welfare (Flink & Kaldewey, 2018; Genus & Stirling, 2018; UNESCAP, 2016). Through various communication channels between firms and society, stakeholders can monitor harmful firm behaviors through transparent corporate governance and stimulate firms to pay attention to business ethics and self-control for corporate social responsibility (Scherer & Voegtlin, 2020) and creating shared values (Porter & Kramer, 2011). After all, the ethical reflexivity of the firm will be useful for RI. Thus, we suggest:

  • Hypothesis 2 The practice of communicative actions toward stakeholders will be positively related to the performance of RI.

Complementary between organizational agility and communicative actions

The interactions between organizational agility and communicative actions can reduce the weakness of each and increase the strength of both, reinforcing each other. First, it can be costly for firms to transform their organizational structure and business operating models into an agile model (Teece et al., 2016). Firms can respond quickly to business environments if they have many agile teams to carry out an agile operating model across the enterprise (Brosseau et al., 2019). Building and launching agile teams requires each team to be accountable for specific results, trusted to work autonomously, enabled to collaborate closely with customers, and capable of fast feedback loops (Rigby et al., 2018). Since communicative actions with stakeholders allow a close relationship with customers and responsibility for management results, firms can meet these conditions to create and implement agile teams. Thus, communicative actions will reinforce organizational agility.

Second, while communicative actions can improve inclusion and reflexivity in RI, it might be difficult for firms to reach an agreement with stakeholders due to conflict between different interest groups. Organizational agility can quickly reduce the complexity of identifying problems and solutions among many stakeholders through experimental and incremental thinking with frequently iterative feedback. In particular, this experimental and incremental thinking is effective in four situations: (1) when the market changes frequently; (2) when customers give quick feedback; (3) when the problems are complex to be solved; and (4) when failures and mistakes can offer valuable learning opportunities rather than catastrophic results (Rigby et al., 2016). Also, agile organizations have organic, informal networks among employees to facilitate open communications across the organization (Brosseau et al., 2019). This organic network in agile organizations can be useful in resolving conflicts with stakeholders through responsive adjustments. Thus, organizational agility can complement the weakness of complexity in inclusive communication actions. In light of this, we suggest the third hypothesis as follows:

  • Hypothesis 3 The RI will improve more in firms that are extensively organized both in organizational agility and in communicative actions toward stakeholders than firms that are not.

Research context and data

The empirical context of the manufacturing industry

We examined the Korean manufacturing industry to test our hypotheses. This industry has been considered one of the major business areas that can generate negative externalities, such as environmental pollution. Regarding business and environmental issues, previous studies have focused on how firms could compete successfully and protect the natural environments simultaneously, that is, the competitive performance implications of corporate efforts with environmental technologies (Berchicci, Dowell, & King, 2012; Berchicci & King, 2007; Rusinko, 2007). In particular, Klassen and Whybark (1999) show that manufacturing firms can enhance both manufacturing and environmental performance by investing in environmental technologies for pollution prevention, management systems for evaluating environmental impacts, and pollution control. Moreover, Pil and Rothenberg (2003) emphasize that corporate efforts to improve superior environmental performance and manufacturing quality are synergistic and reciprocal with each other. Thus, it is important for manufacturing firms to improve the environmental technologies and capabilities for both firm and environmental performance.

Based on this literature, our hypotheses aim to improve our understanding of how firms can achieve environmental performance by linking together organizational factors and innovation activities for green patents. Green patents (i.e., patents in environmental-related technologies) include climate change mitigation technologies, greenhouse gas control, environmental management, and water management technologies (OECD, 2017). These patents are directly associated with the efforts of manufacturing companies to prevent environmental pollution in advance or solve the problems directly ex-post. Therefore, the study of organizational factors and innovation activities in the manufacturing industry can provide us with an appropriate context to understand the underlying mechanism between corporate efforts and RI with green patents.

Sample

We compiled large-scale data, including yearly patent counts, sales portion data, corporate social responsibility data, and financial data, for the attributes of public and privately held manufacturing firms in our sample for the period from 1999 to 2019. We used the TS2000 database developed by the Korea Listed Companies Association (KLCA), a non-profit organization that provides consulting services, such as research reports publishing, accounting/legal system services, and statistical data analysis, with companies listed on the Korean Stock Exchange. The TS2000 database includes publicly audited listed and unlisted firms with assets of greater than 7 billion Won (equivalent to approximately USD 5.2 million at the 2019 exchange rate).

Using Korean industry codes related to manufacturing, we selected 23 manufacturing industries and constructed the panel data of 135,005 observations on 13,234 firms over 21 years to be used in our analysis. According to Table 1, approximately half of our observations in the sample are from six industries related to machinery manufacturing (13.20%), motor vehicles manufacturing (10.54%), electronics manufacturing (10.10%), chemical products manufacturing (8.46%), metal products manufacturing (7.96%), and primary metals manufacturing (7.10%). These heavy industries can generate more pollution than light industries, such as the food and beverage manufacturing industries. In addition, Table 2 presents annual distributions of our observations in the sample. There has been a gradual increase over the years in the number of observations, which may show the continuous growth of manufacturing industries in Korea. That is, manufacturing industries have been more important in the Korean economy over time.

Table 1 Sample: Industry distribution
Full size table
Table 2 Sample: Year distribution
Full size table

Measures

Our dependent variable of Hypotheses 1 through 3 is the performance of RI in manufacturing firms. It can be measured through the number of green patent applications, which represents the patenting activity of each company each year. The count ranges from zero to many or very large for some companies. We focus on the number of patent applications rather than the number of ultimately granted patents to capture the patenting activity regardless of the results awarded by patent officers. From the first econometric model of patent production function using patent data (Hausman, Hall, & Griliches, 1984), researchers have used the number of patent applications to measure the direct intellectual outcomes of scientific efforts (Azoulay, Ding, & Stuart, 2009; Bhaskarabhatla & Hegde, 2014; Hall & Ziedonis, 2001). Thus, we follow this method of measurement to capture the direct outcomes of RI activities.

We collected the number of green patents of the manufacturing firms in the sample using the WIPS ON database developed by WIPS Co., Ltd, the first online worldwide patent information service provider in Korea. WIPS offers the most reliable intellectual property (IP) searching service as the only authorized government-designed prior art search institute by the Korean Intellectual Property Office (KIPO) in the areas of patents, trademarks, copyrights, and industrial designs. We searched for green patents from 13,234 manufacturing companies in our sample, using specific terms, such as ultrafine dust, fine dust, eco-friendly, environmental affinity, and eco. Approximately 1,721 firms among the companies in our sample have 3,849 patents over a 21-year period (1999-2019). Figure 2 presents the distribution of technological classifications (the IPC codes) based on the OECD criteria for green patents. Our sample includes the technology classifications for environmental management, climate change adaptation technologies, and sustainable ocean economy.

Fig. 2
figure 2

The Distribution of Technological Classification in the Sample: OECD Criteria for Green Patents

Full size image

Two independent variables are included in our regression model. The main independent variable is the degree of organizational agility. Although the concept of agility has become pervasive, there has been a wide range of measurements in different contexts and dimensions, as it is difficult to measure agility with a single principle (Yauch, 2011). In this paper, we focus on the organizational agility of manufacturing firms in response to environmental uncertainty. Agile manufacturing (i.e., a quickly responsive manufacturing system) is based on the manufacturer's ability to respond quickly to market requirements, for example, the very rapid launch of multiple new models to market or rapid change of product models in shorter life cycles in response to customers’ preferences (Cachon & Swinney, 2011; Ramesh & Devadasan, 2007). For example, firms with organizational agility can change the manufacturing capacity and reallocate relevant resources quickly across different manufacturing lines to respond to the market requirement for innovation (Kogan, Papanikolaou, Seru, & Stoffman, 2017; Teece, Peteraf, & Leih, 2016) or resilience within global value chains (Scuotto, Garcia-Perez, Nespoli, & Messeni Petruzzelli, 2020). From this perspective, we measure how broadly firms can change the manufacturing ratios in terms of the degree of organizational agility in manufacturing various products in a given time.

We conducted the following procedures to construct the degree of organizational agility in manufacturing companies. First, we used the FnGuide database to collect the sales data for the proportions of the product lines, up to five product lines for each firm. FnGuide, established in 2000, is a privately held financial data analysis and consulting firm in Korea, providing various Korean corporate data and stock information, including around 600 financial companies and government institutions internationally (www.fnguide.com). Then, a firm i’s manufacturing ratios for a product line j in a given time t can be measured as the proportions of sales (PS) for each product line, and the sum of each proportion is 100 (%) as

$$\sum_j{PS}_{ijt}=100,j\in 1,2,3,4,5$$

In the next step, we consider whether the company's manufacturing capabilities are concentrated on a specific product line or expand multiple product lines. The distributed level of manufacturing capacity can be calculated as the mean (μit) and the standard deviations (σit) of the firm i’s PSijt in a year t as follows:

$${\mu}_{it}=\frac{\sum_{j=1}^N{PS}_{ijt}}{N}$$
$${\sigma}_{it}=\sqrt{\frac{\sum_{j=1}^N{\left({PS}_{ijt}-{\mu}_{it}\right)}^2}{N}}$$

In case a company with two product lines focuses on a single product (e.g., 100, 0), the distributed level of manufacturing capacity can be computed as a standard deviation of 70.71.

Finally, we measured the organizational agility (OAit) of the firm i in a year t as the change in manufacturing capacity and resource reallocation. This can be calculated as the absolute value of the changes in the firm i’s standard deviations (σi) between the previous year (t − 1) and the current year (t) as follows:

$${OA}_{it}=\left|{\sigma}_{it}-{\sigma}_{it-1}\right|$$

For example, a firm that reallocates 100% of manufacturing capacity to focus on a specific product line (e.g., 100, 0, 0, 0, 0) in the prior year would get a standard deviation of 44.72. If the firm distributes 20% of manufacturing capacity equally across five lines of products (e.g., 20, 20, 20, 20, 20) in the current year, it will get zero standard deviations. As such, the organizational agility in this manufacturing firm can be |0 − 44.72| = 44.72. Table 3 presents the descriptive statistics for all variables in our sample. Organizational agility ranges between 0 and 70.711, as shown in Table 3.

Table 3 Descriptive statistics
Full size table

Another main independent variable is the practice of communicative actions. We included a binary variable to indicate whether the company published a sustainability report (1 if it was published, 0 otherwise), which means transparent disclosure and communication for stakeholders to understand the firm’s progress toward economic, social, and environmental goals based on the triple bottom line for sustainable development. Most companies in Korea follow the guidelines based on the Global Reporting Initiative (GRI) and ISO 26000 standards to publish their sustainability reports. The publication of sustainability reports is not legally mandatory but is done autonomously in Korea. Global governance for RI, such as the GRI and ISO 26000 standards, can facilitate and enhance innovation activities that avoid harming profits, people, and the planet by leading the central dialogue between firms and stakeholders through sustainability reports (Voegtlin & Scherer, 2017). To collect the data, we used the Korean Sustainability Index database of the Korean Standards Association (KSA), a public organization under the Korean Ministry of Trade, Industry, and Energy.

We included five control variables for important firm-level attributes that may affect the firm’s activities for RI. First, we controlled firm size by taking the natural logarithm of the firm’s total assets. In terms of speed and responsive adaptation, small firms with light assets but well outsourced or networked can carry out dynamic innovation, that is, a series of innovation activities over time, more quickly and efficiently than large firms that have a bureaucratic and vertically integrated organizational structure with heavy assets (Stock, Greis, & Fischer, 2002; Williamson, 1991). Thus, the size of the firm may be negatively related to RI activities. Next, for an equally important reason, we also controlled the number of employees. The third important control variable is firm age. Innovation can be considered a path-dependent activity because scientific efforts and knowledge can accumulate in organizational processes over time (Nonaka & Takeuchi, 1995; Sterman & Wittenberg, 1999). As such, firm age can positively affect the patenting activities (i.e., the number of patent applications) of the firm. Fourth, we also controlled R&D intensity, which can be measured through R&D expenditures nominalized by total assets. In terms of absorptive capacity, investment in R&D activities can be a significant determinant of organizational learning and innovation, assuming that firms can learn and know everything they want (Cohen & Levinthal, 1990). But it can be difficult for an elite firm to solve the complex problems of the grand challenges alone (Brammer et al., 2019). Thus, other stakeholders or ecosystem partners enable the firms to solve the grand challenges together, leveraging their multiple perspectives (Cai et al., 2019). This implies that internal R&D intensity might be less critical relatively in RI activities to solve the grand challenges than traditional innovation activities for product market competition. The other important variable that we controlled for our regression model is the profitability and efficiency of the firm. We measured them by taking the return on assets (ROA). Lastly, we also included the year dummy for each year from 1999 to 2019 to control the time variants and the industry dummy to control the different industrial variants.

Methodology

Our dependent variable is the number of patent applications by a particular firm in a given year. Since this count data is a non-negative integer, we considered applying a Poisson regression model or a negative binomial regression model for our analysis (Hausman et al., 1984). While the Poisson regression model requires that the variance and the mean of the data must be nearly equal, the negative binomial regression model can correct for over-dispersion that the variance is larger than the mean. As shown in Table 3, our dependent variable is strongly skewed to the right. The variance of our dependent variable is .284 (a standard deviation of .533), but the mean is .029, which indicates that the variance is significantly greater than the mean. Since the distribution of our dependent variable presents over-dispersion, we used the negative binomial regression model. Additionally, we examine the zero-inflated negative binomial regression model because approximately 98.7% of our dependent variable is a zero value. The zero-inflated negative binomial regression model can account for excess zeros in the data and over-dispersion (Greene, 1994; Lambert, 1992). Based on the excess zero problem test in a Poisson distribution, the Poisson regression model is appropriate when the observed proportion of zeros in a sample is less than Pr(Y = 0) = e−μ. However, the number of zero values in our sample is almost 98.7%, which is larger than 97.1% (e−.029 = .971). Thus, our sample has an excess zero problem, so the zero-inflated negative binomial regression model is more suitable than the Poisson regression model. Specifically, our empirical model is:

$$Responsible\_ Innovation=\alpha +{\beta}_1 Org\_ Agility+{\beta}_2 Com\_ Actions+{\beta}_3 Org\_ Agility\times Com\_ Actions+{\beta}_4 Employees+{\beta}_5 Size+{\beta}_6 Age+{\beta}_7R\&D\_ Intensity+{\beta}_8 ROA+ Year\_ dummy+ Industry\_ dummy+\varepsilon$$

Findings

Table 3 presents the descriptive statistics. The total number of observations in our sample is 135,005 from 13,234 firms for 21 years. Our dependent variable’s mean (median) value is .029 (.000). The minimum (maximum) number of patent applications is 0 (75). Thus, the distribution of our dependent variables is significantly skewed to the right, as the 75th percentile is still 0. Zero values are still meaningful when analyzing count data, such as the number of patents. This means that the zero-inflated negative binomial regression model is adequate to analyze our count data of green patent applications with excess zero values.

We have two independent variables as organizational factors, such as organizational agility and communicative actions. The average (median) degree of organizational agility is .676 (.000), and the maximum value is 70.711. The communicative actions are measured as a binary variable, 0 or 1. We then constructed the interaction term between organizational agility and communicative actions.

The five control variables capture the unique firm characteristics in our regression model. We controlled the number of employees and firm size to consider the attributes of the organizational structures that could affect continuous innovation over time. The average (median) number of employees is 225.517 (72), which ranges from 1 to 108,998. We take a natural logarithm of total assets to measure the firm size. The firms’ average size (median) is 7.504 (7.375), with a maximum value of 11.340. Firm age is also diverse, from a year to 123 years. Another firm characteristic is R&D intensity, which might be directly related to innovation activities. The average (median) R&D intensity is .009 (.000). It ranges from .000 to 1.286. This shows that firms invest .009 times their total assets on average each year. The biggest investment in R&D represents 1.286 times of total assets. Additionally, ROA is related to the operational efficiency or financial capacity to perform frequent and iterative communicative actions with stakeholders for RI. The average (median) ROA is .035 (.031). It ranges from –21.611 to 8.838.

We report the Pearson correlation coefficient statistics to compare the strength of relationships between two variables in our model. Table 4 presents the correlation matrix for all variables. The significant and positive intercorrelations between organizational factors (i.e., organizational agility and communicative actions) and RI performance are consistent with our theoretical prediction about the effect of organizational factors on RI. In particular, communicative actions are not significantly related to R&D intensity and ROA. This means that there is a trade-off relationship between communicative actions with stakeholders and financial and knowledge-intensive activities at some points of management. Thus, adding communicative actions to our model can give us the opportunity to have some insights into the managerial decision-making issues for RI.

Table 4 Correlation matrix (pearson coefficients)
Full size table

Considering the count data of patent applications, we applied the Poisson, negative binomial, and zero-inflated negative binomial regression models to check the robustness and sensitivity of our findings, using R. Tables 5, 6, and 7 represent Poisson, negative binomial, and zero-inflated negative binomial models respectively. We added the control of the industry dummy and the year dummy in these regression models. In R, the results of nonlinear regression models report deviance values, which indicate the fit of the model. The null deviance shows how well the dependent variable is predicted by a null model that includes only the intercept, while the residual deviation shows how well the model that includes the independent variables predicts the dependent variables. Thus, the residual deviance less than the null deviance indicates a good fit (Cameron & Trivedi, 2005). As shown in Tables 5 and 6, the deviance of the negative binomial model is less than the deviation of the Poisson model, which means that the negative binomial model fits our count data and variables well. Since about 98.7% of our dependent variables are zero, our sample is excessively dispersed, with the data skewed to the right. Therefore, we applied the zero-inflated negative binomial model to analyze our count data as a best-fit model.

Table 5 Effects of organizational factors on the number of green patent applications: Poisson regression model
Full size table
Table 6 Effects of organizational factors on the number of green patent applications: Negative binomial regression model
Full size table
Table 7 Effects of organizational factors on the number of green patent applications: Zero-inflated negative binomial regression model
Full size table

Table 7 presents our main analyses of the effects of organizational factors on the number of green patent applications. Models 1 and 2 are the baseline models to test our main independent variables, that is, organizational agility (Hypothesis 1) and communicative actions (Hypothesis 2), with control variables. Model 3 in Table 7 includes the interaction term of organizational agility and communicative actions and all other variables to test Hypothesis 3. Overall, the signs of the estimated coefficients are consistent with our theoretical predictions, with statistically significant support.

We found strong support for Hypothesis 1. The result suggests that firms that have a higher level of organizational agility perform better RI activities, that is, patenting activities related to green patents. Model 1 in Table 7 indicates organizational agility has a coefficient of .039 and is significant at the .1% level, with a standard error of .002 (z value = 18.810). Holding other variables constant, each unit of the organizational agility level affects the firm’s number of green patent applications by 3.9%. Moreover, the results of Model 2 significantly support Hypothesis 2, which predicts the positive effect of communicative actions on RI, with a coefficient of 1.945 at the .1% significant level and a standard error of .041 (z = 47.953). This is a significant impact compared to the impact of organizational agility. Consistent with the literature on corporate social responsibility and corporate governance for RI (Scherer & Voegtlin, 2020; Voegtlin & Scherer, 2017), communicative actions with stakeholders are the most important factor for firms to drive the orientation of innovations towards RI in response to the grand challenges. We also found that the interaction term of organizational agility and communicative actions has a coefficient of .033 and is significant at the .1% level, with a standard error of .006 (z = 5.514). This finding strongly supports Hypothesis 3, which predicts the positive impact of pursuing both organizational agility and communicative actions on RI. Although we include the interaction term in the model, the coefficients of the other main independent variables still hold to be statistically significant and of similar magnitude with the same signs.

Furthermore, our finding presents that the coefficients of the control variables about firm characteristics, such as firm size, firm age, R&D intensity, and ROA, support our theoretical prediction. The negative signs of the number of employees and firm size present that small firms are better than large firms in RI activities. Consistent with research on organizational structure and innovation from an adaptive perspective (Stock et al., 2002; Teece et al., 2016; Williamson, 1991), it is reasonable to say that small firms are agile enough to respond quickly to environmental uncertainty and grand challenges. However, the small size of the company might not mean newly established firms. This is because firm age positively affects RI, with a coefficient of .005 (a standard error of .002) at the 5% significant level. As innovation is somehow path-dependent, it might be true that old firms may have accumulated more experience and knowledge than young firms. Thus, evidence of better performance in small firms may imply that they may have other external capabilities, such as outsourced partners and network partners in the business ecosystem. Operational efficiency has been shown to have a less significant impact on RI, but R&D intensity has a negative coefficient of –3.413 (a standard error of .439) at the significant level of .1%. We interpret this finding as that the RI in response to the grand challenges does not rely on the sole efforts of one company but could be related to the collective efforts of multiple actors around the company.

Discussion

Theoretical and practical implications

Recent studies on responsible innovation have noted the integration of traditional innovation activities and CSR activities in the context of sustainable growth. The idea is not new that ethical responsibility is an important factor for scientists and engineers to move the edge of scientific discoveries and applications to human society forward. Nevertheless, we still know little about systematically organizing businesses for responsible innovation, which means “taking care of the future through collective stewardship of science and innovation in the present” (Stilgoe et al., 2013). In this study, we advance this literature further by exploring organizational determinants to impact RI performance in response to solving the grand challenges. Theoretically, our model addresses the underlying mechanisms of how organizational responses to grand challenges, such as organizational agility and communicative actions, can affect RI performance implications. The findings offer several important theoretical and practical implications for academics and practitioners regarding RI, the roles of business in solving grand challenges, and agile management.

First, extant research on RI has mainly developed conceptual understanding (Stilgoe et al., 2013) and institutionalization in research and innovation policy (Genus & Iskandarova, 2018; Owen et al., 2021). Yet, what organizational factors drive RI performance has been little known. To fill this gap, we theorize on RI as the process of organizational adaptability in response to solving grand challenges. Our analyses show that the key characteristics of organizational responses, identified as organizational agility and communicative actions with stakeholders, under the emergence of grand challenges have a significant impact on the outcomes of RI. The findings shed light on opening the further extendable research stream of RI by intertwining organizational adaptability and RI performance. In terms of this, our study not only contributes to the literature by identifying critical organizational determinants for the success of RI but also offers practical implications for managers by highlighting the importance of agile management and proactive communications with stakeholders for RI activities. Our empirical evidence supports that managers can enhance the performance of RI if the organizational responses can be systematically manageable in response to any grand challenges.

Second, our study contributes to shifting RI research's attention from ethical responsibility issues to organizational design problems. Prior studies have concentrated on the ethical norms to manage innovation activities and considered RI similar to CSR practices (Genus & Stirling, 2018; Scherer & Voegtlin, 2020; Voegtlin & Scherer, 2017). However, this stream of RI research that simply emphasizes CSR practices may not balance RI theory between the strategic and ethical foundations (Pandza & Ellwood, 2013; Stilgoe et al., 2013). Taking a holistic view of RI, our theorization and analyses include all four foundational dimensions, such as anticipation, responsiveness, inclusion, and reflexivity. We hypothesized the argument that the interaction effect of organizational agility (i.e., internal organizational implementation for anticipation and responsiveness) and communicative actions (i.e., external relationships with stakeholders for inclusion and reflexivity) are likely to improve RI performance across four dimensions and provided empirical evidence. In so doing, our study offers interdisciplinary research implications for scholars to extend the scope of RI research by integrating it with other management areas, such as strategic management, organization theory, and operation management. Our theoretical model may advance our understanding of the underlying mechanisms in turn: organizational responses, strategic implementation processes, and reinforcing the RI foundations to achieve high performance. Moreover, by empirically demonstrating the determinants of RI performance improvement, we support that manufacturing firms might be expected to adopt agile management and inclusive corporate governance actively for sustainable growth with RI.

Third, our findings suggest that not only agile management and stakeholder-based management but also relatively small firm sizes can offer significant implications for any activity within ecosystems to improve RI performance. The strongly supported Hypothesis 3 indicates that the higher the interaction effect of organizational agility and communicative actions, the better the RI performance for relatively small and relatively less R&D-intensive firms. In other words, large firms with high R&D intensity may no longer be an essential indicator for RI in the context of grand challenges. Our results, which are somewhat different from the prediction of traditional innovation research (for example, the advantage of large companies investing heavily in R&D), could have critical implications for RI research. Especially, our analyses imply that there might be some “ecosystem-specific advantages” (Li, Chen, Yi, Mao, & Liao, 2019), for example, open innovation by an interdependently collaborative group of firms to implement RI activities to solve the grand challenges. In general, grand challenges are hard to solve and highly uncertain for huge investments due to societal and organizational constraints (George, Howard-Grenville, Joshi, & Tihanyi, 2016). Hence, large firms are likely to concentrate on scale-up with well-identified solutions through huge investments in traditional innovation for market competition. Unlike the traditional innovation activities of monopolists (i.e., the closed innovation by a large elite firm with huge investment in R&D activities), RI can be viewed as open innovation that requires recombining internal sources of innovation with external sources of innovation from stakeholders within an ecosystem. The main research implication, here, is that the dispersed knowledge of many relevant stakeholders cannot be integrated into one elite firm (Hayek, 1945) but must be turned into collective intelligence through the efforts of all relevant agents with a business ecosystem approach (Cha, 2020).

Overall, we found that the more agile and communicative actions with stakeholders, the higher RI performance in response to grand challenges (e.g., industry innovation and infrastructure, climate action, clean water and sanitation in the UN SDGs) in South Korea. The main practical implications of our findings highlight the importance of agile management of manufacturing processes and stakeholder-based management for RI. In particular, in the Asian-Pacific region, which is highly concentrated on the manufacturing industry, managers must realize that RI performance to solve grand challenges depends on agile and stakeholder-based management. More clearly, managers need to understand that RI activities to avoid negative externalities generating grand challenges differ from traditional innovation activities to build competitive advantages in the product and service markets. As such, managers should reconsider organizational design to be suitable for RI, unlike the traditional approach. Besides, more attention must be paid to agile management than ever before to address management practices in response to grand challenges. In 2020, global value chains and public health management faced the unprecedented risks of the COVID-19 pandemic. We have experienced death, economic lockdown, and other social issues during the pandemic and learned that grand challenges happen unexpectedly and require a rapid response to continuous changes (Gereffi, 2020). Our analyses are related to these management practices in response to the COVID-19 pandemic and empirically support the importance of organizational responses and adaptability for RI.

Limitations and future work

In this article, we tested hypotheses in the context of environmental problems among various grand challenges, such as pollution, water management, global warming, and climate crisis. As we developed the theoretical model of RI with this limited empirical evidence, there are three-fold significant limitations and opportunities for future research.

First, this study examined manufacturing firms in South Korea, where dialogue on environmental issues is pervasive at multiple levels of individuals, businesses, societies, and governments. However, more testing must be conducted in the empirical settings of various grand challenges and different countries. To improve the generalizability of our theoretical model about the impact of organizational factors on RI, researchers can investigate our theoretical prediction in different contexts, for example, the use of artificial intelligence, cities and global development, health and humanity, energy, future of the internet, media, and democracy, etc. (Lufkin, 2017). Thus, by incorporating research on different dimensions of great challenges in society, economy, technology, and environments, future research may reveal other types of organizational determinants or contingent factors that improve RI performance in the different contexts of grand challenges.

Second, in our theoretical model setting, we limited our focus to a basic model with two organizational responses that cause the underlying mechanisms of RI performance. However, from our analyses for testing Hypothesis 3, we acknowledged that there might be other significant interorganizational determinants of RI performance at the ecosystem level. Although we simply noted the ecosystem-specific advantages of small firms for agility and collective intelligence as a research implication, more research is still needed to hypothesize and test whether there are such ecosystem factors in detail. As such, future research can advance our theoretical model by adding more determinants related to the interorganizational dependencies at the ecosystem level.

Third, another limitation of our study is that the dependent variable, the number of patent applications, captures only patentable innovation and technological development. Not all innovations can be patented; for example, designs and software programming codes are documented as trademarks or copyrights. In light of this, future research is needed to overcome the limitation of measuring the outputs of RI in different contexts of grand challenges.

Conclusion

While the literature on RI has been growing increasingly in both conceptual and empirical research (Arslan & Tarakci, 2022; Scherer & Voegtlin, 2020; Scuotto et al., 2022), to our knowledge, this article is one of the first to directly test the impact of organizational factors on the outputs of RI from a holistic perspective, including all four foundational dimensions of RI. Overall, we contribute to the extensive literature in interdisciplinary areas, including strategic management, operations management, governance, and corporate social responsibility, by linking organizational studies with RI for grand challenges. Our analyses also contribute to the literature by developing a theoretical model of underlying mechanisms inside organizations to achieve high RI performance. Clearly, the most important thing is that this study would contribute to developing the holistic approach to the four dimensions of RI in depth with empirical analyses beyond the limitation of current theoretical attention to the meaning of moral responsibility.

In conclusion, as long as manufacturing industry firms pursue their sustainable growth in the context of grand challenges, more future research will benefit from the implications of our RI research.