WORKING PAPERS

​

Quantitative Assessment on Frictions in Technology Market (job market paper) [New version coming soon]

​

In this paper, I first document several novel stylized facts from the Chinese patent transaction data matched with the manufacturing firm data. A key finding is that the Chinese patent market is significantly and consistently less developed than the US. To understand the causes and consequences, I build a patent market model incorporating three types of frictions --- search costs, fixed transaction costs and the information asymmetry on patent quality --- that could shape the size of the patent market. I structurally estimate the model and find the following two main results. First, based on the evaluation of the importance of three frictions calibrated to the Chinese patent market, the elimination of search costs or fixed transaction costs will enlarge the size of the patent market by around 4.5 times, and the productivity growth will increase by 12%-13%. Second, compared with the magnitudes of three types of frictions in the US, search costs turn out to be the main friction to explain the gap of patent market size between China and the US. If search costs in China were reduced to the US level, China’s productivity growth would increase by 7%.

An Anatomy of the Patent Quality: China vs. US, with Hanming Fang, Zheng (Michael) Song and Hanyi Tao [New version coming soon]

​

We construct various quality measures to investigate the evolution of the quality of granted Chinese patents, relative to that of the US patents, over the period of 1985 and 2015. We mainly develop two novel measurement of patent quality that can be comparable across countries: the standardized forward citation by dual-listed patents and machine learning analysis on new words in patents files. Some measures based on network analysis or traditional patent quality measures are investigated as well, like originality and generality, centrality, and patent renewals. Standardized forward citation measure indicates that on average, US patents' quality is two times higher than Chinese; However the quality gap is narrowing, especially in electrical engineering industry which is consistent with the findings in new word analysis, that during 1986-1990, 72.1% of new words that appeared in the Chinese patents were words that already appeared in the US; this fraction continuously dropped, and was 25.3% for the new words that appeared in Chinese patents filed in 2011-2015. In terms of cross-country knowledge transfer as proxied by patent citations and new word citations. We find that the knowledge flowed mostly from the US patents to Chinese patents, though in some technology classes, particularly electric communication technique (H04) and basic electric elements (H01), knowledge started to flow from China to the US from around 2001, though in 2015, the knowledge transfer from China to the US accounted for only 0.91% of the total China-US knowledge transfer.

​