David Hsu

David Hsu and Jeffrey Kuhn (Forthcoming), Academic Stars and Licensing Experience in University Technology Commercialization, Strategic Management Journal.

Suting Hong, Kaihao Guo, Haipeng Zhang, David Hsu (Working), When Do Ventures Founded by Research-oriented Entrepreneurs Outperform?.

Abstract: Holders of a STEM doctoral degree are increasingly pursuing non-academic careers for a variety of reasons. There has been special interest in one such path, founding new ventures, because of its potentially disproportionate impact on societal and economic outcomes. We explore the conditions under which “research-oriented” founders (operationalized as those who have published in the peer-reviewed scientific literature) outperform otherwise observationally similar non-research-oriented founders in venture performance. Our contingency-based inquiry helps reduce tension in the literature which has found seemingly conflicting results on performance advantages to such academic founders. We also extend the human capital construct to the founding team and investor level, as such participants may assist in deficiencies in knowledge and expertise at the individual level. Our data consists of a matched sample of several thousand ventures across a wide span of industrial fields. Our results suggest that research-oriented founders’ ventures outperform when they start ventures in their field of expertise, especially so when they are academically accomplished. Furthermore, for research-oriented founders starting ventures outside of their domain of expertise, recruiting co-founders with complementary functional backgrounds can compensate, though the same statement cannot be said about investors with complementary skills. We end by discussing policy implications.

Matt Marx and David Hsu (2022), Revisiting the Entrepreneurial Commercialization of Science: Evidence from ‘Twin’ Discoveries, Management Science, 68, pp. 1330-1352.

Abstract: What factors shape the commercialization of academic scientific discoveries via startup formation? Prior literature has identified several contributing factors but does not address the fundamental problem that the commercial potential of a nascent discovery is generally unobserved and potentially confounds inference. We construct a sample of over 20,000 “twin” scientific articles, which allows us to hold constant differences in the nature of the advance and more precisely examine characteristics that predict startup commercialization. In this framework, several commonly-accepted factors appear not to influence commercialization. However, we find that teams of academic scientists whose former collaborators include “star” serial entrepreneurs are much more likely to commercialize their own discoveries via startups, as are more interdisciplinary teams of scientists.

David Hsu, Po-Hsuan Hsu, Qifeng Zhao (2021), Rich on Paper? Chinese Firms’ Academic Publications, Patents, and Market Value, Research Policy, 104319.

Abstract: By combining various databases of academic publications and patents of China’s publicly listed firms, we explore the effects of academic publications on firm valuation. We find that Chinese firms’ academic publications are positively associated with their market valuation. More importantly, such a positive relation is more pronounced when these firms have stronger patent records, highlighting a synergy between basic research and applied technologies. Mechanism tests indicate that firms’ academic publications promote their market values through enhancing their human capital and sending credible signals to the market. We also find that publications in English-language journals are more value-relevant than in Chinese-language journals.

David Hsu, Po-Hsuan Hsu, Tong Zhou, Arvids Ziedonis (2021), Benchmarking U.S. University Technology Commercialization Efforts: A New Approach, Research Policy.

Abstract: Despite the economic significance of patented university research, it is difficult to measure the economic value of academic patented inventions and observe the extent to which universities are able to capture such value through patent licensing. Moving beyond assessing commercialization performance by simple statistics, we propose a new approach to benchmarking university patents and commercialization performance based on comparative corporate patent value. Our procedure involves matching university patents to patents with similar patent characteristics granted to public corporations, then estimating the “potential value” of these university patents by stock market reactions to grants of the matched corporate patents. These estimated values of university patents can significantly explain the technology-level income from licensing by a leading US research university and the annual licensing income of the member universities of the Association of University Technology Managers’ (AUTM). We find that AUTM universities realize an average of 16% of the estimated value of matched corporate patents. We also investigate correlates of university-level potential patent value and suggest avenues for future research.

David Hsu, Po-Hsuan Hsu, Kaiguo Zhou, Tong Zhou (Working), Industry-University Collaboration and Commercializing Chinese Corporate Innovation.

Abstract: We examine how industry-university collaboration (IUC) enhances Chinese firms’ commercialization of their technologies using a comprehensive dataset of medium sized and large industrial firms and research universities in China. Measuring IUC with the occurrence and frequency of patents co-assigned to both firms and universities, we first document that firms with more IUC activities report more new product sales. Both dynamic difference-in-differences analysis and instrumental variable regressions support a causal interpretation of the positive relation. In addition, firms with more IUC activities produce more patents and patents that are of higher quality, are closer to basic science, and are more exploratory – all confirming an innovation-enhancing mechanism. Further evidence suggests that our results can be attributed to two channels through which firms benefit from IUC: knowledge acquisition and talent recruiting.

David Hsu, Po-Hsuan Hsu, Yunan Liu, Long Yi (Working), Public Listing, Managerial Short-Termism, and Brand Development.

Abstract: We construct a novel and comprehensive dataset of trademarks owned by big privately held and publicly listed firms in the U.S. over three decades to examine whether and how public listing status affects product inventions. We find that while publicly listed firms register more product inventions compared with their big privately held counterparts, their product inventions are less likely to survive long. We attribute these patterns to the short-termism of public firms’ managers, which is supported by a test based on short-term institutional investors. Using an unexpected court ruling in the Ninth Circuit Court that enhanced managerial short-termism for public firms as an identification test, we find that the relation between the public listing status and product inventions strengthens after this ruling, which supports a causal interpretation of our results.

David Hsu and Prasanna Tambe (Working), Startup Labor Markets and Remote Work: Evidence from Job Applications.

Abstract: Does offering remote work allow startup firms to attract more experienced and more diverse (gender and race) talent? We examine job listings and job applicant behavior on a leading platform in this space, AngelList Talent, amid the COVID-19 pandemic-induced shutdowns. We first characterize the jobs and organizations offering remote work before the shutdowns. We then leverage the context to help address the empirical confound of job design (including offering remote jobs) as co-determined with unobserved job and firm characteristics. By doing so, we estimate the change in applicant characteristics to job postings which are (exogenously) shifted to being remote. This design is a window into evaluating a managerial choice (offering remote work) which will likely become more salient in post-pandemic job design. We find that offering remote-eligible work attracts more experienced and diverse job applicants.

David Hsu and Jeffrey Kuhn (Working), The Outsized Role of Academic Stars in University Technology Licensing.

Abstract: The decline in U.S. corporate R&D over time raises the importance of university-based research in the innovation ecosystem. We examine the process by which university technology transfer offices (TTOs) allocate internal resources, which provides insight into technologies offered for commercialization to the private sector. We compare resource allocation according to academic “star” status against a commercial experience logic. We analyze detailed administrative records of patenting decisions and outcomes by one prominent U.S. research-based university’s TTO over a 30-year period. We find academic stars wield outsized influence: their patent applications are more likely to be filed, yet are less likely to be issued; they are no more likely to have their patents licensed and their patents are no different (statistically) in their licensing revenue from non-stars.

Po-Hsuan Hsu, Dokyun Lee, Prasanna Tambe, David Hsu (Working), Deep Learning, Text, and Patent Valuation.

Abstract: This paper uses deep learning and natural language processing (NLP) methods on the US patent corpus to evaluate their predictive power in estimating two measures of patent value: (i) investor reaction to patent announcements as measured in Kogan et al., 2017 and (ii) forward citations. While forward citations have traditionally been used as measures of economic value in the literature, their utility is mainly retrospective. Contemporaneous predictions of patent value, as embodied in investor reactions to patent grants, can be important for managers and policy-makers for prospective decision making. We compare the prediction performance of models using the structured features of the patent (number of citations, technology class, etc.) to deep learning and NLP methods. Relative to linear regression models using the same features, deep learning models reduce mean absolute error (MAE) by approximately 32%. Incorporating patent text further lowers the MAE by 13%.