Regressions to be run for entire sample of regions and then separately for EU10 group of regions. Estimated using three different econometric methods: two stage least squares (2SLS) and three stage...

Framework and model According to Griliches (1979), innovative input is best reflected by new knowledge, which is primarily embodied in R&D efforts. However, the present model extends beyond them and combines a number of additional variables in the model. The model estimates a cross-sectional knowledge production model over 186 geographic regions in the EU in the period of 1999-2009. R&D output = f (R&D input) (1) The model adopts a general version of the Cobb-Douglas production function, which does not impose any restriction regarding returns to scale, where a is a constant while b measures the elasticity of knowledge output P with respect to input. R&D output = a (R&D input)b exp e (2) Taking the natural logarithm of each side of leads to equation (3): ln (R&D output) = ln a + b ln (R&D input) + e (3) To compare to what extent business and university R&D explain granted patents and innovations in the EU regions, the following model is estimated: ln (P) = ln a1 + a2 ln (BR&D) + a3 ln (UR&D) + a4 ln (HQ) + a5 ln (S&T) + a6 ln (HKIS) + a7 ln (GDP per capita) +e1 (4) Following Jaffe (1989), the potential interaction between university and business R&D is captured by extending the base equation (4) to capture the effects of business and university R&D expenditures, the qualification level of the working age population, number of students at the second stage of tertiary education on parent activity. The structure of the knowledge productive system is proxy by the share of high and medium high-technology industry in total employment. Similarly, it assumed, based on the regional patent data, that there is strong relationship between the numbers of patents application per capita and GDP per capita, e.g. richer regions invest more in patenting activity. As far as innovative output is concerned, it is proxy by patent applications to the EPO, which is the only available harmonized innovation measure at the EU25 NUTS2 regional level. Even if patent data do...