R&D Returns, Spillovers and Firm Incentives: Evidence from China Henry Guofang Huang John Hopkins University Wei Li University of Virginia Lixin Colin Xu World Bank
A Clear and Present Danger: A Non-Randomista is amongst Us! Regressionista
A Clear and Present Danger: Data are almost always Imperfect
China has relied more on Kgrowth lately China has relied on capital growth rather than TFP growth since mid 1990s. TFP growth has played an increasingly smaller role (Zheng and Hu 2006; OECD 2005). From the to the period, – the contribution of TFP for the total growth has declined from 43% to 32%, – that of capital increased from 44% to 62%. Similarly, Lin and Wang (2008): – the share of TFP contribution to GDP growth has dropped from 31% in to 20% in
Rooms from institutional reforms less and less Less from rural migration. Less from ownership restructuring & privatization. Currency appreciation pressure. More has to come from innovation. Similar to OECD and U.S. after 70s oil shocks.
What this paper do Estimate R&D marginal returns How do R&D returns depend on firm incentives & level of development? Within-industry spillover effects? – Allow to differ by distance between firms
Contributions to the study of R&D effects in China substantial R&D returns at firm level in China (Jefferson et al. 2002). R&D inputs increase the share of new products – R&D in China is not merely used as adapting foreign technology in China (Jefferson et al 2002). Returns to private R&D tend to be higher than those to public R&D (Hu 2001). R&D has dropped over time though still substantial (Hu and Jefferson 2004). significant complementarity between own R&D and foreign technology transfer (Hu, Jefferson and Qian 2005). We differ – a random sample, (above-scale large and medium enterprises). – more firm characteristics, – how R&D effects depend on CEO incentives and development level, – allow for the spillover effects and depend on distance between firms, – at a later period when R&D dramatically increased.
Literature two: The general study of R&D effects see Griliches 1998; Mairesse and Mohnen we focus on the largest developing and the fastest-growing economy – instead of samples of firms in developed countries allow return heterogeneity across regions, development levels, and firm incentives. allow the spillover effects to depend on distance between firms. find evidence that indeed spillover effects tend to be smaller for firms located farther apart.
The data The investment climate survey of 120 cities. Cover all provinces except Tibet. 100 firms each city, – Four mega cities: 200 More or less a stratified random sample. General questions to CEOs: cross section Quantitative questions: panel of 3 years. City information
Empirical specifications Estimate the following Interpretation of key coefficients: – Marginal return to R&D capital. – Spillover effects: N firms in industry k * (within-city direct effects + indirect effects for ind k * average distance between firm i and j in ind k)
Conclusions Average R&D return: 0.40 or so. – Reasonably robust. Higher return: – Private firms, corporations (relative to SOEs, collective, foreign). Indications that high costs of capital, higher return: financial constraint? – Non-government-appointed CEO. – No short-term incentives (weak). – Inland areas Significant within-industry spillover – Total spillover effects: additional 1 to 19 percent for some industries. Implications: – R&D returns substantial for a fast-growing economy. – R&D may be one avenue for convergence. – Subsidy for R&D?