Abstract
A central challenge in the measurement of entrepreneurship is accounting for the wide variation in entrepreneurial quality across firms. This paper develops a new approach for estimating entrepreneurial quality by linking the probability of a growth outcome (e.g., achieving an IPO or a significant acquisition) as a function of start-up characteristics observable at or near the time of initial business registration (e.g., the firm name or filing for a trademark/patent). Our approach allows us to characterize entrepreneurial quality at an arbitrary level of geographic granularity (placecasting) and in advance of observing the ultimate growth outcomes associated with any cohort of start-ups (nowcasting). We implement this approach in Massachusetts from 1988-2014, yielding several key findings. First, consistent with Guzman and Stern (2015), we find that a small number of observable start-up characteristics allow us to distinguish the potential for a significant growth outcome: in an out-of-sample test, more than 75% of growth outcomes occur in the top 5% of our estimated quality distribution. Second, we propose two new economic statistics for the measurement of entrepreneurship: the Entrepreneurship Quality Index (EQI) and the Regional Entrepreneurship Cohort Potential Index (RECPI). We use these indices to offer a novel characterization of changes in entrepreneurial quality across space and time. For example, we are able to document changes in entrepreneurial quality leadership between the Route 128 corridor, Cambridge and Boston, as well as more granular assessments that allow us to distinguish variation in average entrepreneurial quality down to the level of individual addresses. Third, we find a high correlation between an index that depends only on information directly observable from business registration records (and so can be calculated on a real-time basis) with an index that allows for a two-year lag that allows the estimate of entrepreneurial quality to incorporate early milestones such as patent or trademark application or being featured in local newspapers. Finally, we find that the most significant gap between our index and the realized growth outcomes of a given cohort seem to be closely related to investment cycles: while the most successful cohort of Massachusetts start-ups was founded in 1995, the year 2000 cohort registered the highest estimated quality.