Discussion

This comparative case study aimed to provide an answer on how managerial cognition in the automotive industry relates to the adoption of inbound-, outbound- and coupled OI activities. The research extends on current cognition literature by looking at the relationship between managerial cognition and the adoption of inbound-, outbound- and coupled OI, and how the relationship is moderated by CEO narcissism.

5.1 Results

Studies have shown that managerial attention is essential for making strategic choices and actions when responding to changing environments (e.g., the four megatrends in the automotive industry), which requires costly and risky investments (e.g., open innovation) (Barr 1998, Cho and Hambrbick, 2006, Kaplan et al. 2003, Ocasio1997). However, little was known about the relationship between managerial cognition and the adoption of inbound-, outbound- and coupled OI specifically. Therefore, this study closed this research gap by building on both cognition and OI literature in a rapidly changing context:

the automotive industry. The cost and innovation pressure in the automotive industry (Ettabaa et al., 2019) is directed by the convergence of the four megatrends (electrification, autonomous driving, connectivity, and the sharing economy) (Accenture, 2018; Gao et al., 2016). The four megatrends force OEMs to look outside their boundaries, and adopt OI activities (De Massis et al., 2012). Bringing this in line with the cognition literature, the assumption was that managerial attention towards the four megatrends would positively affect the adoption of inbound-, outbound-, and coupled OI in the automotive industry. However, the equivalent hypotheses are not supported. On the one hand, the CEO of BMW pays significantly more attention to the four megatrends compared to other OEMs. On the other hand, there is not one OEM who scores considerably higher on inbound-, outbound-, and coupled OI than other OEMs. Therefore, the independent and dependent variables found no relation.

38 However, this research did make a contribution to the OI literature by showing that annual reports are an incomplete source for measuring the adoption of OI activities. Initially, the expectation was that all the inbound-, outbound- and coupled OI activities would be discussed in the annual reports. Annual reports were expected to be a reliable source since it provides applicable information about the past, current and future activities (Breton, 2009). After doing this research, some limitations were found when using annual reports as a source for measuring the activities of inbound-, outbound- and coupled OI.

The annual reports did not always discuss all the OI activities of a firm in a qualitatively. For example, the BMW start-up garage, the venture client of the BMW, was not mentioned in any of the BMW’s annual reports. Considering that this research solely relied on annual reports, the BMW start-up garage is not included in the results. It could be that this limitation only applies to BMW and not to other OEMs, but it does not give a complete picture of the OI activities of BMW anyway. Furthermore, some sentences in the annual reports did not clearly describe the OI activity. For example, in the annual report of BMW, there is the following sentence: ‘’Close collaboration with external business partners and the BMW Group’s own in-house component production team resulted in the introduction of numerous innovations, many of them relating to the BMW i8’’ (annual report BMW Group, 2014, p. 40). After reading this sentence, it was unclear to which type of innovation the collaboration belongs to and how many business partners BMW works with. As a result of the unclarity, it was sometimes hard to categorize the activities correctly. The findings described above show that annual reports are lacking open transparent information, and therefore annual reports give an incomplete view of the adoption of inbound, outbound-, and coupled OI activities. However, the incompleteness of annual reports for measuring the adoption of OI has never been studied before, making it a unique contribution of this study.

Another gap this study aimed to cover was the moderating role of CEO narcissism on the relationship between managerial cognition and the adoption of inbound-, outbound-, and coupled OI. Previous research showed that CEO narcissism positively influences a firm’s performance (Kashmiri et al., 2017;

Mata & Khan, 2020; Zhu & Cheng, 2014). Simultaneously, CEO narcissism proved to enable managerial attention toward new technologies and new developments in a market (Gerstner et al., 2013).

39 Despite these findings, previous research has never studied the moderating effect of CEO narcissism between managerial cognition and the adoption of OI before. Therefore, CEO narcissism was implemented as a moderating factor in this study, and a strengthening role was hypothesized. However, no relationship was found between the independent and the dependent variables. Therefore, the moderating variable did not influence the relationship between managerial cognition and the adoption of inbound-, outbound-, and coupled OI activities. Another explanation for the insignificance of the moderating variable CEO narcissism can be attributed to the measurement method. Where previous researches used a 4-item index (Gerstneret al., 2013; Chatterjee and Hambrick, 2007), this research only uses one of the four items, which could lead to a misrepresented view of the CEO narcissism scores.

5.2 Theoretical contributions

Although the hypotheses are not supported, there were similarities between the literature and the results.

For example, Brunswicker and Chesbrough (2018), Ili and colleagues (2010), Martins and Kaminski (2019), as well as Lazzarotti and colleagues (2013), found that inbound OI activities are more predominance compared to outbound OI within companies. The results of this study confirm this by showing that the OEMs together adopted 88 inbound OI activities in ten years, and 25 outbound OI activities. Coupled OI is a direction that is not discussed often in previous research, but this study shows that all the OEMs together adopted around 77 coupled OI activities. Therefore, the results contribute to the OI literature by showing that inbound OI and coupled OI are both more adopted by OEMs in the automotive industry than outbound OI. The results also confirm Gao and colleagues (2016) that the four megatrends (electrification, autonomous driving, connectivity, and the sharing economy) are revolutionizing today’s automotive businesses. For example, a clear difference between the first three years and the last three years of the managerial attention scores was founded. All four OEMs have doubled their managerial attention scores in the last three years compared to the first three years. This doubling shows that CEOs pay more attention to electrification, autonomous driving, connectivity, and the sharing economy over the years 2010 till 2019. Furthermore, the four trends are also reflected in the inbound-, outbound-, and coupled OI activities. For instance, in the first three years, the OI activities were hardly related to electrification, autonomous driving, connectivity, and the sharing economy.

40 Whereas, in the last three years, a big part of the inbound-, outbound- and coupled OI activities are related to the four megatrends.

5.3 Implications

Current research shows that the automotive industry is experiencing a revolutionary discontinuity driven by four megatrends electrification, autonomous driving, connectivity, and the sharing economy. In times of a changing industry, managerial attention is important in order to respond to the change (which could result in adopting OI activities). Therefore, this research tried to examine the relation between managerial cognition and the adoption of inbound-, outbound-, and OI activities. Although the relationship was not supported, this research has implications for other firms in incumbent industries.

The findings show that over the years, more inbound and coupled OI activities are related to the four megatrends in the automotive industry. Due to the changing industry, OEMs are forced to make new investments, such as the investment in electrification. New investments go along with high risks and costs. To spread these risks, OEMs joined forces that resulted in joint ventures or strategic alliances (Reid, 2016; Deloitte, 2012). Firms in other incumbent industries experiencing a revolutionary character could therefore adopt inbound- and coupled OI as well, to spread risks in new investments. Take for example the banking industry, which is currently undergoing a transformation and where there is uncertainty of what the future will look like (Smith and Eckenrode, 2016). New startups are disrupting the traditional financial systems by enhancing the current financial services and widening the consumers’ choices (Boustani, 2020). To acquire relevant technologies for banks and stay competitive in the market, OI should be considered as a good opportunity to spread risks in these new investments.

A contribution of this research is that annual reports are an incomplete source for measuring the adoption of OI activities. If firms want to get to know all the OI activities of all their competitors or other firms, they should not solely rely on the annual reports. The implication is that annual reports are lacking open transparent information. For a complete picture of the OI activities, annual reports can be used in combination with other sources such as websites, press releases, and interviews with employees.

However, firms or researchers can still use this study as a guideline for measuring the adoption of

41 inbound-, outbound-, and coupled OI. For example, they can look for the OI activities as mentioned in tables 6, 8, and 10 by using multiple resources. Another example for further research could be that the inbound-, outbound-, and coupled OI activities of companies are analyzed via critical media literacy, which consists of analyzing mass media and popular culture practices such as radio, TV video, and newsletters (Alvermann and Hagood, 2000).

5.4 Limitations and future research

One limitation was that a comparative case study was chosen as no previous studies researched OI in the automotive industry based on annual reports. As a result, there were no clear guidelines for measuring OI in annual reports in the automotive industry yet. To first find out what the annual reports about OI contain, an intensive investigation of annual reports was needed. Unfortunately, this investigation was a time-consuming process, and therefore the sample size needed to stay small.

However, this explorative investigation contribute to the OI literature as it was the first of its kind to study OI in a qualitatively with annual reports as a source. Future research should be conducted with a larger sample size to see whether it generates similar results.

While the annual reports did not include all the OEM’s OI activities, the missing OI activities in the annual reports could affect the final results. Therefore, further research is needed with different types of sources to map all the inbound-, outbound-, and coupled OI activities. For example, sources like websites, press releases, and interview with employees. Furthermore, the cases (BMW, FCA, Ford, and Tata) belong to the top fifteen biggest car manufacturers and although a pre-test was done on the independent variable and moderating variable to ensure some difference between the chosen cases, the difference between the amount of inbound-, outbound- and coupled OI was not significant. The difference might be more considerable once they are compared to the results with two OEMs who are not in the top fifteen biggest car manufacturers. Therefore, future research could implement OEMs in the sample that do belong to the top fifteen biggest car manufacturers.

Furthermore, content analysis was done manually to gather the data of all the variables. For the independent variable, manual content analysis was chosen to ensure that all relevant words that were

42 counted were related to the four megatrends and not to something else. However, a limitation is that the researcher could overlook words that could directly affect the results. To further ensure reliability, an extra researcher was asked to measure the same sample of LTS’s. The adoption of inbound-, outbound- and coupled OI was measured via a manual content analysis as well. This process was done by looking for specific words and activities which go along with an intensive investigation of the annual reports.

Here, a limitation again is that the researcher could overlook activities due to inconsistency or tiredness.

However, the manual approach was needed since there was no clear guideline on how to measure the adoption of inbound-, outbound-, and coupled OI before. Further research could perform the same process with the help of automatically screening qualitative data analysis software.