Data for: How patent signals affect venture capital: the evidence of bio-pharmaceutical start-ups in China

Published: 21-05-2019| Version 1 | DOI: 10.17632/623sknzjth.1
Ying Guo,
lili Zhang,
Ganlu Sun


To conduct our analysis, we built a dataset comprising 457 bio-pharmaceutical start-ups in China and their patents. The start-up VC data was sourced from the Zero2IPO database, a comprehensive, accurate and timely professional database covering VC events since 1992 in China . It is a reliable data source from one of China’s major large venture capital data information service providers and many research about China’s venture capital used it (Dang et al., 2011; Luo et al., 2016). We got the data on 3 November 2017 using the industry classifications ‘biotechnology’ and ‘medicine’ as the search criteria. Information on specific VC investments was gathered using a web crawler. Only investments in the angel, seed, Pre-A, and A~G rounds was captured, along with the “New Third Board” rounds. (The “New Third Board” is an OTC equity financing scheme sponsored by the Chinese government that allows listed companies to issue a non-public offering of stocks to specific targets with some restrictions). We used Tianyancha and Qichacha (two comprehensive enterprise information query tools, widely used in China’s business practice and academic research (Lili et al., 2018; Xizi & Wenyuan, 2017)) to collect other basic information about each start-up, e.g., date of registration, registered address, etc. We excluded any records with incomplete fields, resulting in a total of 535 VC investments into 457 start-ups (some start-ups received multiple investments). Patent data was retrieved from PSS using the name of each startup as the search word under ‘Patentee’. First, we searched using the full name. If that search returned no results, we searched again using a selected keyword in the name. Retrieved patents were only recorded and counted if the patent was granted prior to a VC investment. What’s more, there might be difference between out-licensed and in-licensed patents in the view of investors. So we just chose the start-ups whose patents have no record of patent transfer. Of the 457 start-ups, 227 held existing patents at the time of financing and these firms received 275 investments. However, there are only two VC events are invented by state-owned VC, the Sdicfund. It’s too few to control this variable. So we delete these two events to make sure all the VC investors are private VC, which means we get 226 start-ups and 273 events with patents.