The Globalization of R&D: China, India, and the Rise of International Co-invention Lee Branstetter, CMU and NBER Guangwei Li, CMU Francisco Veloso, Catolica, CMU 1
In conventional models, innovative capability takes time to develop Per capita GDP Complex manufacturing Light manufacturing Agriculture New technology creation Time (Vernon, 1966; Stigltz,1970;Grossman & Helpman, 1989) 2
China and India are still at an early stage in the development process $46,860 GDP per capita 2010 $42,783 $40,274 $20,756 $18,558 $10,816 $9,522 $7,274 $4,382 $1,371 United States Japan Germany Korea Taiwan Brazil Mexico South Africa China India 3
But China and India are already innovating Total number of USPTO patent grants 4
Experts have pointed to the rise of innovation in China and India as a potential challenge for our models and for the U.S. and other advanced economies Daniel Trefler and Diego Puga (2010) Wake up and smell the ginseng! Richard Freeman (2006) Rapid convergence of China/India to U.S. patterns of comparative advantage could create new competition for U.S. high-tech exports Paul Samuelson (2004) The same textbook models that illustrate the benefits of U.S. trade with China and India also illustrate how those benefits could vanish if the convergence in comparative advantage is complete 5
A careful look at patents granted by the USPTO to inventors based in China 6
China and India differ from Taiwan and Korea in the composition of their innovation surges 7
The innovation surges in China and India are driven by MNCs Majority of patents are owned by MNCs About half are co-invented The important role of MNCs in the international invention explosions of China and India may help explain why they are occurring at a early stage of economic development. 8
Agenda Using U.S. patents to map out innovations in China Insights from a field study Regression analysis of China-generated U.S. patents 9
Chinese domestic patent data suggest an explosion of innovation 10
But the numbers of true patent grants (invention patents) are much smaller 11
A significant fraction of Chinese invention patents are awarded to foreign inventors 12
US firms aggressively patent their inventions in other major markets 30 25 20 15 10 U.S. applicants 5 0 U.S. Europe Japan Two Triadic Markets Three Triadic Markets 13
But the top 100 indigenous Chinese applicants patent only a small fraction of their inventions outside China 30 25 20 15 10 5 U.S. applicants Chinese applicants 0 U.S. Europe Japan Two Triadic Markets Three Triadic Markets 14
We are not the only ones who question the value of Chinese patent grants Brian Wright and his students have found that Chinese indigenous inventors inflate their patent applications to meet local government targets And to benefit from local government subsidies Domestic patents of low quality can also be an asset in an evolving legal system that struggles to distinguish between a good patent and a bad patent The number and growth rate of domestic patenting may (substantially) overstate the true innovation of indigenous Chinese firms 15
Breakdown of USPTO patents assigned to Chinese inventors by assignee nationality and technology field Assignee Nationality Technology Field 16
Chinas Special Innovation Zones Beijing Greater Shanghai Guangdong 17
Agenda Using U.S. patents to map out innovations in China Insights from field study Regression analysis of China-generated U.S. patents 18
Interviews of China-based (and India-based) multinational R&D personnel Time: December 2009 Locations: Beijing, Shanghai Main interviewees: researchers or directors including inventors listed on U.S. patents from six multinational R&D subsidiaries 4 research facilities affiliated with U.S.firms 1 industrial university cooperative research center fully sponsored by a Taiwanese firm 1 European chemicals and pharmaceuticals firm Discussions with China-based academic experts Telephone conversations with high-level multinational managers in China and India. 19
Lessons from interviews International co-invention is a real phenomenon, not just an artifact of patent data MNCs see China- (and India-) based engineers as an essential resource for their global R&D operations MNC operations in China (and India) are re-engineering Western products for local markets AND creating new technology for global markets 20
Agenda Using U.S. patents to map out innovations in China Insights from field study Regression analysis of China-generated U.S. patents 21
Are co-invented / multinational sponsored patents are of higher quality within China? Model: Poisson Quasi Maximum Likelihood Estimation C is the total number of non-self citations China-based patent i receives as of the end of 2010. Coinv, MNC are dummies indicating the patent was created by an international team and/or assigned to a multinational PatStock is the assignee f s three-year patent stock before the date of application. TeamSize is the total # of inventors on the patent. H = technology fixed effects and T = time fixed effects. Baseline of comparison: USPTO patents owned by indigenous firms 22
Cross-firm Comparison within China 23
How does team composition affect the quality of patents produced by the same multinational in different countries? Coinv, Domestic are dummies denoting whether or not a patent is created through co-invention or by a purely Chinese team. f denotes firm. F controls for firm fixed effects. Baseline of comparison: USPTO patents created by inventor teams entirely resident in the MNC s home country 24
Cross-border Comparisons within MNCs (China) Consider only patents assigned to MNCs from 1996-2009. Match with comparable patents created solely by inventors in MNC home county. Same firm assignee, three-digit technological class, grant year. 25
Over time, the relative quality of MNCs China-generated patents has increased substantially 26
How does the invention quality gap between MNCs and indigenous firms change over time? Divide data into 3 periods according to grant year: <2000 2000-2004 2005-2009 Model: add interaction terms to cross-firm model 27
China: the gap b/w co-invention and purely Chinese invention is disappearing; the gap b/w MNCs and Indigenous firms remains 28
Observations Within the border, across firms Co-invented patents are of higher quality (as measured by forward citations) than patents created by indigenous inventors, both for China and India. Patents under the sponsorship of MNCs are of higher quality than those under the sponsorship of indigenous enterprises in China. A patent quality gap b/w MNCs and indigenous firms persists in China and India, a quality gap b/w co-invented and indigenous patents persists in India Within the same MNC, across borders Aggregating over time, co-invented and purely Chinese-invented patents are equivalent in quality to patents generated in the home country. China-generated patents have risen sharply in relative quality over time India-generated patents are of lower quality than patents generated in the home country. Evidence of relative quality improvement in India is weaker than in China. 29
Conclusions and implications Our research suggests that R&D processes can now be disaggregated into multiple stages, which are then located where they can be undertaken most effectively. China s (and India s) rise in international patenting surge is driven largely by multinational firms based in advanced economies. Simply counting Chinese patent grants significantly exaggerates real innovation by indigenous Chinese firms. Can China and India breathe new life into the Western innovation machine? Jones (2009) suggests diminishing productivity in R&D as the burden of knowledge rises. But this can be offset by plugging enough new scientists into a globalized innovation process. 30
Thank you! branstet@cmu.edu guangwei@cmu.edu 31
Cross-firm Comparison within India 32
Cross-border Comparisons within MNCs (India) Consider only patents assigned to MNCs from 1996-2009. Match with comparable patents created solely by inventors in MNC home county. Same firm assignee, three-digit technological class, grant year. 33
The relative quality of MNCs India-produced patents has not grown like their China-produced patents 34
India: the gap b/w co-invention and purely Indian invention persists 35
And this helps drag patent renewal rates in China well below those of any other major patent jurisdiction 36
Lessons from interviews (1) Vertical specialization exists in the R&D process (Krugman, 1995; Hummels, Ishli & Yi, 2001; Yi, 2003); there is an international division of research labor The division of labor is not always simply high-end in the West, low-end in China (or India) The main reason for going to China (or India) is to tap the local talent pool Re-engineering products for the Chinese market (and the Indian market) is a source of co-invention Modern telecommunications technologies play a critical role in international R&D collaboration 37
Lessons from interviews (2) MNC executives tended to regard the innovative capacity of indigenous firms as well behind that of the leading MNCs This was confirmed by an indigenous Chinese engineer who left a multinational to work at a local firm then came back! The managers and engineers working in the local firms are quite capable But the iron law of comparative advantage leads them to focus on being low cost manufacturers rather than product innovators 38