Time series and panel data analysis of GEDI and growth performance indicators

N.S. Bosma, J. Content, M.W.J.L. Sanders, F.C. Stam

Research output: Book/ReportReportProfessional


In this paper we look into the hypothesis that institutions drive high quality entrepreneurship that in turn promotes innovative and inclusive growth. We start our analysis by estimating baseline growth models for 25 EU countries over the period 2006-2014. We compare our results to those of growth model specifications in the tradition of Islam (1995) and Caselli et al.(1996) and then include various measures of entrepreneurial activity in these growth regressions. In addition, we control for, and interact with, institutional variables that we show to have a direct positive effect on our entrepreneurship variables. Regulation of credit, labour and business positively affects (high quality) entrepreneurship, while the size of the government is negatively linked to entrepreneurship. For these institutions, that have shown to promote entrepreneurship we apply a GMM panel data estimation method to account for endogeneity following the recommendations of Bjørnskov and Foss (2016). We find a positive link between entrepreneurship and GDP growth. These effects are reinforced after controlling for institutional characteristics that are linked to entrepreneurship in the existing body of literature. Our report thereby strengthens the empirical foundation under a more integrated and institutionalist approach to the study of entrepreneurship. We discuss extensively how this report relates to the work on the Global Entrepreneurship Index in earlier FIRES-reports. In our conclusion, we discuss various possible explanations for our results and offer implications for institutional reform agendas aimed at stimulating entrepreneurship.
Original languageEnglish
PublisherAcademic Press
Publication statusPublished - 2017


Dive into the research topics of 'Time series and panel data analysis of GEDI and growth performance indicators'. Together they form a unique fingerprint.

Cite this