Abstract
ABSTRACTPast literature has used conventional spatial autoregressive panel data models to relate patent production output to knowledge production inputs. However, research conducted on regional innovation systems points to regional disparities in both regions’ ability to turn their knowledge inputs into innovation and to access external knowledge. Applying a heterogeneous coefficients spatial autoregressive panel model, we estimate region-specific knowledge production functions (KPFs) for 94 NUTS-3 regions in France using a panel covering 21 years from 1988 to 2008 and four high-technology industries. A great deal of regional heterogeneity in the KPF relationship exists across regions, providing new insights regarding spatial spillin and spillout effects between regions.
Highlights
Aquaro, Bailey and Pesaran (2015) make the observation that space-time panel data samples covering longer time spans are becoming increasingly prevalent
EuroLIO report to the Federation Nationale des Agences d’ Urbanisme, 2016. It has important implications in terms of regional public policies, standard knowledge production functions do not allow us to identify if some regions are less likely to benefit from internal and external knowledge or to generate spillout effects
Our heterogeneous coefficients spatial autoregressive panel model is capable of producing estimates of knowledge production function parameters for each region in the sample
Summary
Aquaro, Bailey and Pesaran (2015) make the observation that space-time panel data samples covering longer time spans are becoming increasingly prevalent. They clearly stress heterogeneity in regional effects of knowledge generation and diffusion, all these studies fail to provide a coherent region-specific investigation This is because their knowledge production function estimates are based on homogeneous coefficients, assuming that the ability to generate and benefit from spatial knowledge diffusion is the same for all regions or groups of regions. Conventional spatial econometric investigations require a large group of (contiguous) regions to produce reasonable estimates of the role played by spatial dependence/interaction between regions This leads to conventional panel data models, that produce a set of parameters describing the relationship between the different explanatory variables and the dependent variable based on N regions over T time periods. Since our Bayesian estimates are likelihood-based, they share the same consistent asymptotic properties as the QML estimates of Aquaro, Bailey and Pesaran (2015) in cases when normal priors distributions are centered on the true parameters, or where prior variances of normal prior distributions approach infinity
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have