Linking national business system with firm level innovation: The mediating role of intellectual capital and absorptive capacity

As business becomes more fast-paced in the early of 21st century, organizational success relies ever increasingly on innovation. Firm's capability to innovate, however can be influenced by a range factors that reside at different analytical levels. In recent years, institutionalists are increasingly trying to explore the impact of macro-level institutions on micro-level (i.e. firm) phenomena. A large body of research in this field, however, has failed to delineate clear mechanisms to explain such linkages. Therefore, this study develops an integrated model with the aim to explore the link between national business system and innovation output at firm level through examining the mediating role of intellectual capital and absorptive capacity in a developing country context. To test the research hypotheses, this study has followed positivist philosophical research tradition and has applied quantitative research methodology. This approach statistically tests theories to determine effects or outcomes of the study and facilitates to generalize such outcomes to the overall population. The research design of this study bears the qualities of the cross-sectional survey research. Target respondents of the survey were set as primarily owner/ managers and in case that this criterion was not met; senior managers and managers were set as secondary respondents. This study followed snowball-sampling method consistent with the loosely coupled sampling frame and key informants methodology. In order to consider a firm as a part of sample, this study has set two relatively simple criteria. First, it was specified that sample firms should be driven from knowledge intensive industries in order to observe a variance in the dependent variable, the innovation output. Second, firms to be included in the sample should be driven by the population of private and local firms in order to concentrate our attention in dominant economic actors of the Pakistani setting. Prior to the data collection, current study engages in an extensive scale development effort to collect representative data about the national institutions and business system of the Pakistani setting, complemented by relevant validity and reliability tests. To do so, this study followed established guidelines and close attention was paid to item wording, structure of the statements, item meaning, item scaling, and appropriate numbers of items so that respondents can understand the statements without any difficulty and confusion which normally result from 'double barreled' items. Likewise, scaling is an important aspect in scale development activity. The research espoused a Likert-type scale that is the best known and widely used in behavioral research tradition. Next, this study has created an initial pool of items that would tap into each construct's domain and considering the centrality of societal institutions and business systems scales, their items, and other constructs' items were reviewed by subject matter experts (SMEs) to assess the face and content validity. Prior to this, SMEs were briefed about the conceptualizations of all constructs. Based on their feedback, all items of each construct were included in the initial scale and sent to fifteen businesses executives for further identification of any ambiguity or difficulty that might hamper filling the questionnaire out. Furthermore, they were asked to provide suggestions for improvement or any other changes that they deemed necessary. Minor improvements were suggested in wordings of several items, which were adjusted after consulting with the SMEs. After developing a preliminary version of the scales including cover letter and guidelines, it was decided to launch a pilot study to ensure the reliability and usability of scales. The objective of pretesting was to minimize those errors that can be problematic during final data collection, and these errors generally become apparent due to a poor research design or use of ambiguous wording in the questionnaire. Out of 60 distributed questionnaires, 35 filled questionnaires were received back. Two were having substantial missing data, with the omission of these two, remaining 33 were of use. After reviewing the comments of respondents regarding item wording, sentence structure, and level of difficulty or confusion in understanding the language used in the questionnaire, few suggestions about changing some words and sentence structure for several items, which we incorporated in the final survey. Additionally, exploratory factor analysis (EFA) was carried out to examine the factor structure and internal consistency (reliability) of the scales. The coefficient alpha for each construct achieved the threshold value of 0.70. However, it is suggested that these results should be treated with extra caution because EFA needs relatively larger sample size, particularly in the case of unclear factor pattern. Accordingly, these results were considered as indicative of main study data patterns, rather than final. Therefore, no items were deleted merely based on these findings. Target population of this study is firm from knowledge intensive industries including textile, pharmaceutical, engineering, information technology, electrical/electronics, and automotive/automobile. In order to reduce potential variation caused by the macro-level factors, extensive effort was put to collect data in a short interval between the months of April, 2016 and October, 2016. Research assistants, who were postgraduate students, collected the data with face-to-face and on-site visits. They were adequately trained about the purpose of the research and theoretical nature of constructs prior to the data collection. Researchers made visits to the companies with one month interval after the initial visit in cases that the data were not collected within the first round. Out of 1,235 distributed questionnaires, 228 questionnaires collected back with 18.46% response rate. Out of these, 14 questionnaires dropped out due to a large amount of missing data, thus 214 usable questionnaires were included for further data analysis. Before testing the hypotheses of study, several data cleaning methods such as missing data analysis, outliers' detection and assessment of normality were applied to obtain precise results. After cleaning the data, exploratory and confirmatory factor analyses were applied. Lastly, reliability and validity of each scale was tested. The results of all these tests were found satisfactory to proceed the main analysis of the study. The overall theoretical model was tested by structural equations using partial least squares approach. This approach is suitable for small sample size studies and does not require data normality as needed in covariance-based SEM approaches. The results of fit and quality indices of the model achieve the set criteria, thus confirming a high predictive power of the model. The results of path coefficients of the relationship between latent variables were obtained after controlling firm age, firm size, export, R&D budget, firm type (family or non-family), and affiliation with business group. Firm size significantly influences both the radical and incremental innovation. These results are obtained after controlling firm age, firm size, export, R&D budget, firm type (family or non-family), and affiliation with business group. Almost, all variables have nonsignificant impact on both types of innovation output, except business group application on incremental innovation. In addition to this, R&D budget and firm size have insignificant negative impact on radical innovation. Export has insignificant negative impact on both radical and incremental innovation output. The results exhibit strong support for the mediating role of repository (intellectual capital) and enabling (absorptive capacity) cognitive factors in the association of macro and meso level institutional characteristics and firm-level innovation. Thus, this study significantly extends the literature about the national business systems approach by providing a clear-cut mechanism about how abstract institutional templates constituted at the societal and business system level are translated into actionable sets by the help of aforementioned cognitive factors. Moreover, hypotheses on direct links between the study constructs were also supported, thereby indicating variables reside at different analytical levels are closely linked in a topdown manner. In addition, the particular characteristics of the Pakistani setting also reflect themselves in the empirical results. First, the fragmented and informal nature of the dominant societal institutions necessitated the use of non-ownership based control and coordination mechanisms alongside with ownership based control and coordination mechanisms. Thus, it can be argued that the owners pragmatically resort to any means necessary to control and coordinate the economic resources of their firms by engaging in relationships with a variety of stakeholders including the state, competitors, employees, financial organizations and the community. Although, theoretical discussions in the NBS approach suggest that there are only few possible NBS typologies because of the interdependencies between different institutional templates, the Pakistani setting seems to exhibit the characteristics of a hybrid model. Pakistan's business model generally displays relatively higher levels of direct ownership control and coordination characteristics, lower levels of non-ownership based coordination and control characteristics, and relatively tighter levels of interdependence between employer-employee relations. The implication of present research is to contribute institutional theory and to elaborate how NBS theory could be used in developing countries. It particularly demonstrates how the Pakistani business system influences firm level innovation outputs and offers guidelines for management in designing successful innovation related policies and practices in Pakistan. Moreover, this study pinpoints for managers how knowledge management activities inside the firms and grasping outside knowledge can enhance innovation output that can lead them to stay ahead of competitors. Although constraints exist in the form of lack of resources or access to the resources, yet firms can manage such constraints by developing network ties be more innovative.