December 5, 2014
Roman Lubynsky, MIT; Erin Scott, NUS Business School; and Pian Shu, Harvard University
Scott, Shu, and Lubynsky provide some of the first clean empirical evidence that it is possible to assess the commercial viability of a venture even in the idea stage. Collaborating with the Massachusetts Institute of Technology's Venture Mentoring Service (VMS), the researchers examine detailed data on 655 early stage venture ideas across a wide range of industry sectors. Venture ideas that receive a more positive evaluation from a large set of skilled practitioners in the New England entrepreneurship and technology community (i.e., the mentors at VMS) are significantly more likely to ultimately reach commercialization with recurring revenue and expenses in the future. This relationship is economically meaningful and statistically significant after controlling for a rich set of venture characteristics. Due to the educational nature and institutional structure of VMS, initial mentor evaluation is unlikely to affect the degree of mentoring a venture receives nor the behavior of the founding team. Econometrically, the authors also control for mentor-venture interactions and present several robustness checks demonstrating that the intensity of mentoring is not driving their key findings. The evaluations in their setting are based on standardized, concise, and objective venture summaries prepared by VMS staff and focus primarily on the underlying business proposition. Mentors have not met the teams at the time of evaluation. Thus, the authors' results suggest that the business proposition of a venture, even in a very early stage, has some signal value. Leveraging the richness of their data, the authors further show that such signal value is relatively strong for ventures that focus on developing new technologies but weak for ventures that focus on developing new business models with existing technologies. These results have important implications for understanding the risks of entrepreneurship and the allocation of entrepreneurial resources.
Deepak Hegde, New York University, and
Justin Tumlinson, University of Munich
Why do individuals become entrepreneurs? When do they succeed? Hegde and Tumlinson develop a model in which individuals use pedigree (e.g., educational qualifications) as a signal to convince employers of their unobserved ability. However, this signal is imperfect, and individuals who correctly believe their ability is greater than their pedigree conveys to employers, choose entrepreneurship. Since ability, not pedigree, matters for productivity, entrepreneurs earn more than employees of the same pedigree. The authors' preliminary empirical analysis of two separate nationally representative longitudinal samples of individuals residing in the U.S. and the U.K. supports the model's predictions that (A) Entrepreneurs have higher ability than employees of the same pedigree, (B) Employees have better pedigree than entrepreneurs of the same ability, and (C) Entrepreneurs earn more, on average, than employees of the same pedigree, and their earnings display higher variance. The researchers discuss the implications of their findings for entrepreneurship, education, and public policy.
Juanita Gonzalez-Uribe, London School of Economics, and Michael Leatherbee, Stanford University
Gonzalez-Uribe and Leatherbee estimate the effect of business accelerators on start-up performance. The researchers focus on the case of Start-Up Chile (SUP), an accelerator sponsored by the Chilean government, which provides 40,000 USD (equity free) in seed capital, a work visa, and free office space to 100 participants every 4 months. The authors collect information on all start-up applicants since inception of the programme in 2010, and track their subsequent performance using web-based metrics such as presence in fund raising sites like AngelList and Crunchbase, and social-media activity in Facebook and LinkedIn. They use the original ranking of applications by external judges, together with the discrete jump in the probability of selection around the 100-th (ranked) company, as an instrumental variable for selection into SUP. Their results do not allow the authors to rule out the possibility that participation in SUP has no impact on subsequent start-up performance. They contemplate several explanations for the findings, and argue that the most likely explanation is that there is heterogeneity in the effect of SUP, and that the average quality of start-ups near the 100-th company threshold is too low for them to benefit from the accelerator services. Using data from the "the Highway", an accelerator within the accelerator only available to top-performing participants, the researchers find stronger evidence that business accelerators positively impact start-up performance.
Entrepreneurship and Job Creation
Manuel Adelino and Song Ma, Duke University, and
David Robinson, Duke University and NBER
New firms are an important source of job creation in the economy, but the mechanisms underlying the link between new firms and employment growth are not well understood. This paper focuses on employment creation as a result of new investment opportunities and asks whether growth is driven by new firms or by the expansion of existing firms. Adelino, Ma, and Robinson use regional industrial structure and national changes in manufacturing employment to identify shocks to local income, and examine employment creation in the non-tradable sector. New firm entry is much more responsive to changing local economic conditions than growth by established firms. Moreover, their responsiveness doubles in areas with better access to small business finance. Although the researchers focus on the non-tradable sector for identification, their results extend to the construction sector and the economy as a whole, indicating that the mechanisms they uncover are economically pervasive.
R&D and Entrepreneurship
Annamaria Conti, Georgia Institute of Technology
Government R&D grants are widespread policy instruments to ease startups' liquidity constraints. These grants generate spillovers that are appropriable by foreign agents. Small open economies are the most concerned and often impose restrictions on the offshoring of government-funded output, increasing foreign investors' opportunity costs. Examining Israeli startups, Conti finds that these restrictions act as screening mechanisms inducing startups to reveal their characteristics. Ex-ante, startups with a high probability of being acquired by foreign companies are deterred from applying for grants. Ex-post, grant funds positively affect the probability that their recipients experience a successful exit but not the probability that foreign companies acquire them. Finally, grant recipients receive follow-on financing but not foreign venture capital.
Yael Hochberg, Rice University and NBER;
Carlos Serrano, Universitat Pompeu Fabra and NBER; and
Rosemarie Ziedonis, University of Oregon
The use of debt to finance risky entrepreneurial-firm projects is rife with informational and contracting problems. Nonetheless, Hochberg, Serrano, and Ziedonis document widespread lending to startups in three innovation-intensive sectors and in early stages of development. At odds with claims that the secondary patent market is too illiquid to shape debt financing, the researchers find that intensified patent trading increases the annual rate of startup lending, particularly for startups with more redeployable (less firm-specific) patent assets. Exploiting differences in venture capital (VC) fundraising cycles and a negative capital-supply shock in early 2000, the authors also find that the credibility of VC commitments to refinance and grow fledgling companies is vital for such lending. Their study illuminates friction-reducing mechanisms in the market for venture lending, a surprisingly active but opaque arena for innovation financing, and tests central tenets of contract theory.