Entrepreneurship is thought to be a key driver of economic growth. While there are myriad forms of entrepreneurship, ranging from self-employment to small and medium size enterprises to technology- and innovation-driven startups, recent research provides evidence that the relationship between entrepreneurship and economic growth is driven not by overall quantity of new firm entry, but rather by a small subset of high-growth startups that are primarily categorized as innovation-driven. This paper provides a survey of the growing literature on the economics of such innovation-driven entrepreneurship. We begin by distinguishing between the various forms of entrepreneurship, which are often confounded in both theory and empirical work. We lay out the current state of knowledge, and describe the challenges faced by researchers in the field, particularly around measurement, data and identification. We conclude with an overview of the major open questions and directions for future research in the area.

We use a randomized experiment with 553 science- and technology-based startups in 12 co-working spaces across the US to evaluate the effects of intensive, short-term entrepreneurial training programs on survival and performance for innovation-driven startups. Treated startups are more likely to shut down their businesses and do so sooner than control startups. Conditional on survival, however, treated startups are more likely to raise external funding for their ventures, raise funding faster, and raise more funding than the control group; they also exhibit higher employment and revenue. Treated founders are less likely to found a new startup after shutdown. Our findings are consistent with practitioner arguments that early entrepreneurship training interventions can help entrepreneurs with less viable ventures “rationally quit” (“fail fast”). We use machine learning techniques (causal random forest) to provide exploratory insights on the most impacted subgroups. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Using a regime change in the commercialization of university innovation in 1980 that strongly increased university incentives to patent and license discoveries, we document that an increase in the supply of commercializable innovation attracts venture capital investment to the region. The Bayh-Dole Act shifted ownership of intellectual property stemming from federally-funded research from the federal government to universities, spurring technology transfer into the local area. Because universities have different technological strengths, each local area surrounding a university experienced an increase after 1980 in commercializable innovation relevant to particular sets of industries which differed widely across university counties. Comparing industries within a county that were more versus less related to the local university's innovative strengths, we show that venture capital dollars after 1980 flowed systematically towards geographic areas and industries affected most by the sudden influx of commercializable innovation from universities. These results persist even when controlling for ex ante geographic and industry distributions of corporate patenting and prior venture financing. The findings support the notion that increased supply of commercializable innovation serves to draw private capital investment to a region.

Using panel data on 251,511 patent inventors matched with voter registration records containing partisan affiliation, we provide the first large-scale look into the partisanship of American inventors. We document that the modal inventor is Republican and that the partisan composition of inventors has changed in ways that are not reflective of partisan affiliation trends amongst the broader population. We then show that the partisan affiliation of inventors is associated with technological invention related to guns and climate change, two issue areas associated with partisan divide. These findings suggest that inventor partisanship may have implications for the direction of inventive activity.

Startup accelerators, which aim to improve the set of choices representing a startup's entry strategy, have become increasingly influential in both regional development and the strategies of individual startups. This article explores an accelerator's impact on startup performance and whether that impact varies substantially by features of the startup's founding environment. Leveraging data from a leading startup accelerator, I use a regression discontinuity framework to hold startup quality constant so that I can compare the performance of admitted startups to those that do not make the cut, and I examine whether any observed performance differentials are driven by accelerator admission and by characteristics of the startup's earlier environment. I find evidence that startups from better pre-accelerator environments experience stronger gains from accelerator admission. I also find evidence of home bias, as local startups have a stronger treatment effect. These results provide evidence of ecosystem effects whereby the impact of one organizational sponsor in an ecosystem is strongly moderated by other features in the ecosystem. The findings help to explain the concentration of accelerator programs in already successful entrepreneurial ecosystems and reveal how such programs may interact with founding environments to complement resource abundance or magnify prior resource inequalities.

This study investigates the gender gap in entrepreneurship in the technology industry. Digitization has created vast economic opportunities in the technology sector and has lowered many barriers to entry, thus reducing traditional frictions regarding entrepre-neurship and potentially increasing opportunities for female founders. However, anecdotal evidence has suggested that female technology founders are rare and that women are underrepresented in science, technology, engineering, and mathematics roles. Based on individual career histories collected from more than 42 million U.S.-based LinkedIn profiles, including more than 1.3 million founders, we explore whether there are higher rates of female founders in technology companies relative to other industries. Our analysis revealed the following: (1) Females were only half as likely as males to found businesses in the tech-nology industry. (2) Females were less likely to found successive businesses (i.e., serial founders), which was even more pronounced in the technology industry. (3) When we used the gender gap in labor force participation as a baseline, the gender gap in technology entre-preneurship was particularly large, even compared with other male-dominated industries (e.g., construction). (4) The gender gap in technology entrepreneurship was driven by lower rates of entrepreneurship by females in lower positions in the organizational hierarchy. In contrast, females who reached the C-suite in technology sectors were 16% more likely to found firms compared with their female C-suite counterparts in nontechnology industries. Combined, the results provide a nuanced view of the gender gap in entrepreneurship.

Patents are key strategic resources which enable firms to appropriate innovation returns and prevent rival imitation. Patent examiners – individuals who may be subject to various sources of bias – play a central role in determining which inventions are awarded patent rights. Using a novel dataset, we explore if one increasingly prevalent source of bias – political ideology – manifests in examiner decision-making. Reassuringly, our analysis suggests that the political ideology of patent examiners is largely unrelated to patent office outcomes. However, we do find evidence suggesting politically active conservative-leaning examiners are more likely to grant patents relative to politically active liberal-leaning examiners, but only for patent applications where there is ambiguity regarding what constitutes patentable subject matter and hence examiners have greater discretion.

While universities are key sites for the development of new and innovative ideas, translating basic research into commercial products and companies is not so straightforward. The frequency and extent of commercialization of basic research findings varies significantly across universities. In this paper, we examine how commercialization of university-based research depends on the region in which the university sits. We merge data from the Association of University Technology Managers on licensing, spinoffs, and revenues with features local to the university, such as industry mix and availability of entrepreneurial finance. We find that being in a city significantly increases university commercialization outcomes, even after accounting for a robust slate of controls.

Research Summary We introduce a new database which provides an unprecedented window into the off-the-job lives and interests of public firm top executives as reflected by their personal income allocation. We construct this database by matching household credit card spending data with the population of executives in Execucomp. To overcome the significant computational challenges associated with matching these data, we build on the statistical record-linking literature in a way that allows us to generate reliable matches with limited information. To facilitate research exploring new questions made possible with this database, we make our matching crosswalk freely available for academic use. Managerial Summary This article describes the matching procedures associated with the development of a database which sheds new light on the revealed preferences of public firm top executives. Prior work on upper echelons has routinely stressed how preferences and characteristics of top executives often manifest in firm behaviors and are important predictors of firm outcomes. Nevertheless, the lack of a reliable paper trail capturing executive preferences, particularly at a large scale, has been a friction slowing comprehensive empirical research on this topic. We describe how we address this limitation by linking top executives listed in the database Execucomp with credit card spending data provided by the consumer data provider L2 and outline a number of new research questions made possible by our matching efforts. The results of our matching efforts (Execucomp-L2 unique identifier crosswalk) are available at .

Does local innovation attract venture capital? Using a regime change in the commercialization of university innovation in 1980 that strongly increased university incentives to patent and license discoveries, we document the complement to Kortum and Lerner (2000)’s finding that financing leads to future innovation. Because universities have different technological strengths, each local area surrounding a university experienced an increase in innovation relevant to particular sets of industries after 1980—industries which differ widely across university counties. Comparing industries within a county that were more versus less related to the local university’s innovative strengths, we show that venture capital dollars after 1980 flowed systematically towards geographic areas and industries with the greatest sudden influx of innovation from universities. In contrast, the geographic and industry distribution of corporate patenting and prior venture financing in the pre-period does not predict a differential increase in future venture financing, suggesting that our findings are not solely driven by the 1979 pension fund reform that increased financing available to VCs across the board. The results support the notion of a “virtuous cycle” wherein innovation serves to draw capital investment that then funds future innovation.

Background. A meta-analysis of trials in endovascular therapy suggested an increased mortality associated with treatment exposure to paclitaxel. Multiple publications and corrections of prior data were performed, and the United States Food and Drug Administration has issued multiple advisories regarding paclitaxel use. We analyzed how this controversy impacted device purchasing and related utilization patterns in the period immediately following publication of the meta-analysis. Methods and Results. Ascension Healthcare System purchase data over a 14-month period were synthesized across centers for both paclitaxel and non-paclitaxel devices. A fixed-effects regression model and a binary regression model with facility-level controls were used to compare purchasing patterns before and after the meta-analysis. Purchase volumes of each paclitaxel device fell. Pooled purchase volumes of all paclitaxel devices decreased from a 14-month peak of 631 devices in October 2018 to a 14-month nadir of 359 devices in February 2019. An F-test comparing the pooled-month specific fixed effects for the months before vs after the publication of the meta-analysis has an F-statistic of 11.64, suggesting that average purchasing levels in the two periods are statistically different (P<.001). Utilization of non-paclitaxel devices did not decline. Conclusions. Purchase volumes of paclitaxel devices decreased immediately during the months following publication of the related meta-analysis. Total Ascension-wide paclitaxel device purchase volume in February 2019 demonstrated a 43.1% reduction from peak monthly purchase volume during the assessed period and a 32.5% reduction compared with November 2019, the last month preceding publication of the meta-analysis.

Accelerator programs are an increasingly important part of entrepreneurial ecosystems. While accelerators have core defining features-fixed-term, cohort-based educational and mentorship programs for startups- there is also significant variation amongst them. In this paper, we relate key variation in the antecedents, organizational design and operation of these programs to theories of firm-level entrepreneurial performance. We then document descriptive correlations between these design elements and the performance of the startups that attend these programs. In doing so, we probe the connections between design and performance in ways that integrate previously disparate research on accelerators and expand our understanding of startup intermediaries. Our findings delineate the building blocks as well as an agenda for future researchers to build upon not only our understanding of accelerators, but also our understanding of what new ventures need to survive and flourish.

We use electronics badges to measure in-person communication across companies from an accelerator program, and analyze its relationship with their performance. Our analysis shows that both subjective and objective performance correlates with the amount communication exhibited by early stage companies. In general, more communication correlates with better performance, though too much communication with other teams seems harmful. Lower internal communication entropy correlates with higher performance. Companies that spent more time with the program mentors do better. Large companies reported higher levels of satisfaction compared to small companies.

We examine the spillover effects of seed accelerator programs—fixed-term, cohort-based educational and mentorship programs for startups—on venture-backed seed and early stage technology startup activity in the regions in which they locate. Accelerators are often opened with an eye towards galvanizing technology entrepreneurship activity in the region through provision of peer effects/role models. We use a difference-in-differences approach that utilizes the staggered introduction of such programs combined with matching methods and synthetic control methodologies to assess the impact of an accelerator’s arrival on the volume of seed and early stage VC deals completed in the region, excluding the accelerator’s own portfolio companies. The arrival of an accelerator is associated with a significant increase in the volume of seed and early stage deals external to the accelerator cohorts; an increase driven both by outside investor groups and the emergence of new local early-stage investors, and supporting the notion that an accelerator can lead to peer effects and provision of role models in the ecosystem that encourage additional local entrepreneurial spillover activity. The findings suggest that the introduction of such programs can have a general effect on the equilibrium of the regions in which they locate, rather than merely an effect of treatment on the treated, and suggests a role for accelerator programs in galvanizing latent regional interest in entrepreneurial activity.

Combining data from the social progress index and measures of economic institutions and performance, our analysis focuses on how changes in economic institutions and performance are related to subsequent changes in social progress (noneconomic dimensions of societal performance). We document a positive relationship between improved economic performance and subsequent social progress improvements, a separate impact of improved economic institutions on aspects of social progress that involve individual investment (such as education and health), and a noisy relationship between economic factors and those aspects of social progress related to issues of individual freedom and social inclusion.

Despite their increasing importance for both sides, we understand little about the process generating alliances between startups and existing large firms. Two separate theoretical channels presente...