PLS based Analysis on Innovation Effects in Small-medium-sized Enterprises of China-Comparative Study between Native and Foreign Owned Enterprises

摘要

Innovation effects, as one of the research focus in international innovation policy studies, are more and more frequently targeted on developing countries, such as countries and districts in Asia, and particularly on Chinese situations, due to the increasingly crucial role of innovation and technology resource in these increasingly competitive markets, especially when there is multinational enterprises and overseas invested firms in the market.This study, by applying Partial Least-Squares Regression (PLS regression) to the innovation effect issue, focuses on monetary input, human resource input, organizational input, and their related outcomes in small-medium-sized enterprises. The methodology may provide better solution to avoid multi-co linearity of variables in the study. Besides the application of PLS methodology, unique parts of the study can be summarized into two folds: 1)three kinds of resource input cover major components for innovation system, including human resource, research and development activities (reflected through realized intra company expenditures), and institution factors; 2)Comparative study is applied to test whether the native and foreign owned enterprises should pay attention to the same kind of input resource or not. And patent data are applied as outcome indicator. It typically reflects innovation nature in developing countries, typically in Chinese industries.Major findings in the first part include, human resource input and monetary input are more significantly influencing innovation outputs, while numbers of research institutions play less role in innovation activities of small-medium-sized enterprises. And major findings in the second part indicate that there are rather bigger differences between the two groups.Based on these careful investigations, influencing degree of those three input factors is also examined and compared, and the study confirms the patent data can be used as more convincing factors for related studies. Finally, PLS is proved to be more effective methods to provide more robust findings.