Three Empirical Essays on Innovation and R&;D Spillover

• Main Authors: Liao, Kuei-Yen, 廖貴燕
• Other Authors: Chyi, Yih-Luan
• Format: Others
• Language: en_US
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/74718491722545049685

博士 === 國立清華大學 === 經濟學系 === 100 === This dissertation studies the relations between innovation and R&;D spillover. While the spatial variations and productivity differences have been largely attributed to dense knowledge flows and local labor concentration, the explanatory power of R&;D spillover on firm innovation presents ambiguously and no consensus has been reached to date. To explore at greater depth, my research rigorously investigates what lies behind the R&;D spillover on innovation of clustered firms. Substantial progress has been achieved in unraveling the complicated composition of R&;D spillover on firm performances. It contains of three essays. In the first essay, “Innovation, Clustering, and Knowledge Spillovers: An Empirical Study of Finnish and Swedish Firms,” I assess two questions. One is that whether high-productivity firms sorting into industrial clusters, can benefit from agglomeration and thus advance their innovation. The second one is that whether, controlling for firm sorting, R&;D spillover is vital in explaining firm patenting. Due to the patent data containing excess zeros, I apply two Zero-inflated count data models to a panel of 147 Swedish and Finnish firms during 2002-2005. Firm size, productivity, R&;D stock intensity and regional human capital contribute mostly in boosting firm patenting. After controlling for firm sorting effects, industrial clustering may not be overwhelmingly effective across industries in initiating patenting by located firms. The cross-cluster R&;D spillover positively affects patenting intensity, while the effects of within-cluster R&;D spillover vary substantially among industries. The positive influences of local R&;D spillover on firm patent applications are limited to specific industries. To dig deeper into the regional R&;D spillover, further efforts are made on disentangling different R&;D spillover effects. The second essay “R&;D Spillover and Patent Production in British, French and German Firms” thereby investigates the impacts of R&;D spillover across regions and industries on the innovation performance of top R&;D investing firms in France, Germany, and the United Kingdom. Based on a theoretical framework in Bloom et al. (2012), two new measures of technological spillover and product market competition spillover are constructed by using R&;D investment and patent data for 606 firms and 76 regions during years 2002-2007. Application of the non-linear GMM panel count method to solve the endogeneity of RD input and patent production, the results showed that technological spillovers positively associate with patenting by French and German firms, whilst French firms have experienced negative business stealing effects from product market rivals. Especially, the encouraging effect of regional R&;D spillover experienced by German firms is substantial in stimulating located firm patenting due to the dominating beneficial technological spillover effect over the unfavorable spillover effect driven by product market competition. The third essay, “Policy-induced Agglomeration, R&;D Spillover and Firm Performance,” I revisit how a policy-induced science park experiment on firm performance is affected by firm-selection, or by agglomeration economies, or a combined effect of the both. The Heckman selection model of the PSM method is applied to avoid sampling selection bias arising from observable differences between the 128 target Taiwanese stock-listed electronics firms locating in science parks and the 600 non-target firms placed elsewhere. My findings show that the target firms persistently outperform those non-target ones in terms of total factor productivity (TFP) efficiency and patent applications during 2002-2006 due to a combined effect of agglomeration economies and of self-selection by high performance firms. The system GMM panel dynamic models were established to lessen the endogeneity and performance dynamics problems in the examination of the second hypothesis assuming that the effects of technology spillover and local labor density positively affect performances of target firms. The results partially support the second supposition, especially in that firms concentrated in an area with the prevalence of technology spillover and higher labor density are best suited for reaping agglomeration benefits in innovation. As performance feedbacks reinforce agglomeration premium of clustered firms; this in turn, drives more high-performance firms to science parks. My findings have important policy implications. Regarding whether policy efforts should encourage more clusters, the results are mixed. In Nordic countries, firms in three out of eleven industrial clusters have fewer patents than those in isolated areas, whereas firms locating in science parks outperform those placed elsewhere in Taiwan. While the regional R&;D spillover effect on firm innovation is ambiguous, the beneficial technology spillover effect overwhelmingly boosts firm patenting across different regions and industries in France, Germany and Taiwan. A country or region can improve its innovation performance by encouraging spatial agglomeration only when the firm-region interaction of R&;D activities is implemented in the close technology fields. Otherwise, a pro-cluster subsidy simply attracts the less efficient firms to relocate leading to no gains from agglomeration. Keywords: Innovation; R&;D Spillover; Patent; Technology Spillover; Product Market Competition; Agglomeration; Science Park; TFP; Labor Density