«John R. Graham, Co-Supervisor David T. Robinson, Co-Supervisor Manuel Adelino Alon P. Brav Manju Puri Aaron K. Chatterji Dissertation submitted in ...»
The Life Cycle of Corporate Venture Capital
Department of Business Administration
John R. Graham, Co-Supervisor
David T. Robinson, Co-Supervisor
Alon P. Brav
Aaron K. Chatterji
Dissertation submitted in partial fulﬁllment of the requirements for the degree of
Doctor of Philosophy in the Department of Business Administration
in the Graduate School of Duke University 2016 Abstract The Life Cycle of Corporate Venture Capital by Song Ma Department of Business Administration Duke University
John R. Graham, Co-Supervisor David T. Robinson, Co-Supervisor Manuel Adelino Alon P. Brav Manju Puri Aaron K. Chatterji An
of a dissertation submitted in partial fulﬁllment of the requirements for the degree of Doctor of Philosophy in the Department of Business Administration in the Graduate School of Duke University 2016 Copyright c 2016 by Song Ma All rights reserved except the rights granted by the Creative Commons Attribution-Noncommercial Licence Abstract This paper establishes the life-cycle dynamics of Corporate Venture Capital (CVC) to explore the information acquisition role of CVC investment in the process of corporate innovation. I exploit an identiﬁcation strategy that allows me to isolate exogenous shocks to a ﬁrm’s ability to innovate. Using this strategy, I ﬁrst ﬁnd that the CVC life cycle typically begins following a period of deteriorated corporate innovation and increasingly valuable external information, lending support to the hypothesis that ﬁrms conduct CVC investment to acquire information and innovation knowledge from startups. Building on this analysis, I show that CVCs acquire information by investing in companies with similar technological focus but have a diﬀerent knowledge base.
Following CVC investment, parent ﬁrms internalize the newly acquired knowledge into internal R&D and external acquisition decisions. Human capital renewal, such as hiring inventors who can integrate new innovation knowledge, is integral in this step.
The CVC life cycle lasts about four years, terminating as innovation in the parent ﬁrm rebounds. These ﬁndings shed new light on discussions about ﬁrm boundaries, managing innovation, and corporate information choices.
5.2 Direct Information Acquisition from Portfolio Companies....... 49
5.3 Integration of CVC-Acquired Information through External Acquisitions 52
I would like to express my sincere gratitude to all those who helped me complete this dissertation. First, I am indebted to my dissertation committee: John Graham (co-chair), David Robinson (co-chair), Manuel Adelino, Alon Brav, Manju Puri, and Ronnie Chatterji. John is my role model as a researcher, teacher, and adviser, as well as an active contributor to the profession, and I thank him for always being there for me and shaping who I am as a scholar. Likewise, I am extremely grateful to David, not only for his guidance on exploring my research agenda, but also for the chance to learn under him as a co-author. Manju opened the door of entrepreneurial ﬁnance research to me and provided crucial feedbacks at each stage of my dissertation study. Alon taught me a great deal, through patient conversations as well as heated debates, on how to produce rigorous ﬁnance research and be a critical thinker. Manuel provides expert advice on my papers, my job market, and how to establish myself as a young scholar. I am also grateful to Ronnie for encouragements and his unique view of CVC research.
I would also like to thank a number of faculty who were not on my committee but who nonetheless were very gracious with their time and eﬀort—Hengjie Ai, Ravi Bansal, Simon Gervais, Cam Harvey, Wei Jiang, Adriano Rampini, and many others.
I also thank the participants of many workshops that I presented at for valuable feedback that has signiﬁcantly improved this work. I also learned a great deal and immensely enjoyed the close friendship of many fellow students: Gabriel Bonilla, xii Sergio Correia, Shiming Fu, Donghee Jo, Ben McCartney, Tae-Rog Oh, Alexandru Rosoiu, Basil Williams, and Yichong Zhang,.
My family has made many sacriﬁces (beginning long before I enrolled at Duke University) and oﬀered me a tremendous amount of support, and for that I am truly grateful. My parents, Suhui Yang and Frank Ma, taught me to follow my dream and have been extremely supportive of the path I chose. Their constant pride, if at times undeserved, has nevertheless been a source of inspiration to me. I owe a tremendous thanks to my parents-in-law, Hua Wei and Peng Zhao. Without their assistance over the past year, it would have been physically impossible for me to go on the job market and defend my dissertation.
All of this is second to the friendship, love, and support (and occasional prodding) of my wife, Yize Zhao. Ever since we fell in love in college, she has been my inspiration and motivation for continuing to improve my knowledge and move my career forward.
She is my rock. I hope that over the course of my life I am able to repay the love and steadfast devotion she has shown me. I also want to express my thanks to my daughter, Carol, for always being such a good girl and cheering me up.
IntroductionOne important theme in economics is to understand how to make investment, organizational, and ﬁnancial decisions to achieve innovation and long-term growth.1 Financial economists seek to understand the economics of ﬁnancing innovation by examining such activities as internal research and development, mergers and acquisitions, and strategic alliances.2 Yet the pathways to successful innovation and long-term growth remain far from clear (Lerner, 2012).
This paper contributes to this broad research agenda by studying ﬁnancing innovation through the lens of Corporate Venture Capital (CVC), an emerging yet understudied piece in the innovation puzzle. As opposed to producing innovation internally through research and development (R&D) or purchasing external innovation through mergers and acquisitions (M&A), corporations create Corporate Venture 1 See, e.g., Schumpeter (1942), Arrow (1962), Aghion and Tirole (1994), and Klette and Kortum (2002). This topic is crucial to understanding such broad economic questions as economic growth and productivity (Acemoglu, Akcigit, Bloom, and Kerr, 2013; Aghion, Akcigit, and Howitt, 2014;
Bloom, Schankerman, and Van Reenen, 2013), ﬁnance (Brown, Fazzari, and Petersen, 2009; Hall and Lerner, 2010), entrepreneurship (Chemmanur, Loutskina, and Tian, 2014; Acemoglu and Cao, 2015), and organizational economics (Fulghieri and Sevilir, 2009; Hackbarth et al., 2014).
2 See, e.g., Robinson (2008), Brown et al. (2009), Manso (2011), Bena and Li (2014), Seru (2014), among others.
1 Capital (CVC) divisions to make systematic minority equity investments in early stage entrepreneurial ventures. Consider, for example, GM Ventures, the CVC unit initiated by General Motors in 2010. On behalf of General Motors, GM Ventures invested in dozens of auto-related technological startups, including automotive cleantech, advance materials, among other ﬁelds, through minority equity stakes. The case of GM Ventures is hardly an isolated occurrence. According to the National Venture Capital Association (NVCA), CVC investments accounted for about 20% of Venture Capital (VC) investment in 2015,3 and are undertaken not only by technology ﬁrms in the media spotlight (such as Google Venture and Intel Capital) but also by moderate-size ﬁrms in a variety of industries.
Firms state that they conduct CVC investment to acquire information—that is, to gain exposure to new technologies and markets which in turn beneﬁt their capacity for innovation and broader corporate decisions (Siegel et al., 1988; Macmillan et al., 2008).4 As GM Ventures states, its mission as a CVC is to “invest in growth-stage companies to enhance GM’s ability to innovate.” Theoretically, this information acquisition rationale echoes a long-held framework in the economics of managing and organizing innovation that dates back at least to Nelson (1982).5 The innovation process is typically framed as a two-stage sequential process, in which ﬁrms acquire information and generate ideas (ﬁrst stage) before they invest in and produce those ideas (second stage). Including this information acquisition stage in studying innovation can reconcile many important patterns in economic growth and innovation dynamics 3 In 1999 and 2000, CVC investment peaked at almost $20 billion a year. Although the numbers throughout the VC industry fell in the 2000s, CVC growth has rebounded in recent years (37% increase in 2015, 69% increase in 2014, according to NVCA).
4 Startups are an important source of technological and market knowledge, as well as innovative ideas (Scherer, 1965; Acs and Audretsch, 1988; Kortum and Lerner, 2000; Zingales, 2000).
5 See also Nelson and Winter (1982); Dosi (1988); Fleming and Sorenson (2004); Frydman and Papanikolaou (2015), among others.
2 (Jovanovic and Rob, 1989; Kortum, 1997).6 Despite the importance of acquiring information, little empirical work has studied how ﬁrms organize their investment to achieve this goal. Focusing on CVC therefore provides a unique empirical setting to understand how ﬁrms seek out and generate new ideas in the broad innovation process.
Motivated by the empirical signiﬁcance and value of CVC for understanding the process of innovation, I develop empirical tests to assess its role in acquiring information. Those tests examine each stage of a CVC, from why it is initiated, to how it is operated, to when it is terminated. The results establish an investment pattern, labeled as the CVC life cycle, that shows how CVC ﬁts into the process of corporate innovation—deteriorating ﬁrms in need of information launch CVC; they actively invest in and use valuable information and knowledge gained through CVC; and they terminate CVC after having regained an edge in innovation therefore informational beneﬁt shrinks—lending support to the information acquisition rationale. Using these ﬁndings, I further explore alternative forces that could determine CVC investment and the management of innovation in general, as well as the implication of CVC on corporate innovation at both the ﬁrm and industry levels. Figure 1.1 summarizes the CVC life cycle.
The CVC life cycle begins with the initiation stage, in which a ﬁrm launches CVC investment, typically following a deterioration in internal innovation. That is, when ﬁrms experience a decline in internal innovation, and therefore can beneﬁt from 6 Existing studies have overwhelmingly focused on the second stage of the innovation process— investing and organizing innovation with an exogenous idea and predetermined informational structure. Aghion and Tirole (1994) model several cases in which, taking the research idea and informational environment as given, equity investment is optimal to provide incentive for R&D projects;
Mathews (2006) and Fulghieri and Sevilir (2009) study the problem of strategic equity investment from the industrial organization perspective, and theorize the beneﬁts of coordinating market entry and obtaining competitive advantages; Hellmann (2002) emphasizes that asset complementarity and product market synergies lead ﬁrms to invest in synergistic entrepreneurial ventures, particularly when external ﬁnancing is costly (Allen and Phillips, 2000).
3 potential informational gains by connecting to highly innovative entrepreneurs, they are more likely to initiate CVC investment. Quantitatively, a two-standard-deviation decline in innovation quantity (quality) increases the probability that a ﬁrm will initiate CVC by about 52% (67%). Firms in the same industry cluster CVC activities when their sector experiences technological shocks, forming “industry CVC waves.” To mitigate the concern that innovation capability is endogenously determined, I identify plausibly exogenous shocks to ﬁrms’ ability to generate ideas and produce innovation. The instrumental variable Knowledge Obsolescence captures the evolution of usefulness of a ﬁrm’s knowledge base accumulated in the past, which results from exogenous technological evolution but is independent to such endogenous factors as corporate governance and product market status. In addition, several potential alternative interpretations of the result, such as the eﬀects of ﬁnancial constraints and excess cash, do not explain this ﬁnding.
4 5 Figure 1.1: The Life Cycle of Corporate Venture Capital How do CVCs select and use information and innovative knowledge? At the operation stage of the CVC life cycle, I examine how CVC units strategically select which portfolio companies to acquire information from. I ﬁnd that CVCs primarily invest in startups that are innovating in technological areas that are close to the CVC parent ﬁrm, suggesting that CVCs prefer to invest in companies whose technologies can be adapted to the parent ﬁrms’ innovation. Moreover, the portfolio companies appear to possess incremental knowledge, measured using fewer overlaps of innovation proﬁles and patent citations, which suggests that CVC parents aim to acquire updated knowledge with higher informational gain.7 This means, for example, that an automobile CVC parent ﬁrm is likely to invest in an engine startup, particularly when this startup specializes in cutting-edge clean-tech that the outdated parent ﬁrm does not possess.
I perform two analyses to isolate how CVC-acquired information is incorporated into parent ﬁrms. First, CVC parent ﬁrms appear to internalize acquired knowledge through research incorporating the information acquired from their portfolio companies. Second, they capitalize on information by gaining eﬃciency when making information-sensitive decisions, such as external acquisitions of companies and innovations. Moreover, human capital renewal, such as hiring additional inventors who can use the newly acquired knowledge, is integral to this information acquisition and integration.