Causal dynamic effects in regional systems of technological activities: a SVAR approach

Abstract

This paper analyzes the causal relationships in regional technological systems within a structural vector autoregression framework. Applying a data-driven independent component analysis, it shows how the regional dynamics of economic, research, innovation and educational activities affect each other instantaneously and over time. By matching differently classified data on employees, patents and graduates, the analysis is based on a unique database which embraces multi-dimensional aspects for five clearly separated industries. The findings on how industry specific growth processes unfold are explained by referring to the type of industry and its knowledge base. For instance, it is found that more engineering-oriented industries like machine tools show a success-driven pattern, whereas science-based industries show a stronger dependence on universities. Such knowledge on the causal relations is of utmost relevance to the design and implementation of policy instruments. JEL Classification

The effects of public research and subsidies on regional structural strength

“…The exclusion of cyclical immediate dependencies is called the acyclicity assumption and might not be given in regional development. As a test for this assumption, the robustness of the causal ordering is used (see Brenner and Duschl 2015 for details) and we find sufficient robustness (74%).…”

“…The VAR-LiNGAM approach has already been applied in a number of studies (e. g. Brenner et al 2018;Brenner and Duschl 2015;Coad et al 2012). In these studies, median regressions (least absolute deviation (LAD) method) are used within the VAR model (for details see Brenner and Duschl 2015).…”

“…The VAR-LiNGAM approach has already been applied in a number of studies (e. g. Brenner et al 2018;Brenner and Duschl 2015;Coad et al 2012). In these studies, median regressions (least absolute deviation (LAD) method) are used within the VAR model (for details see Brenner and Duschl 2015). In our application, heteroscedasticity is an issue and therefore we apply a log-linear regression by using the logarithm of all included variables.…”

Regional effects of university funding: Excellence at the cost of regional disparity?

“…We use the classification of German labor market regions from the scheme for "improving regional economic structures", which is commonly used in regional research on Germany (see e. g. Brenner and Duschl 2015). Specifically, our analysis is based on the classification for the year 2007, which shows the best fit with our overall observation period.…”