Modelling the green knowledge production function with large panel data econometrics

This study estimates an aggregate green knowledge production function (GKPF) for 19 OECD countries from 1981 to 2012, using panel-data econometric methods to address spatial spillovers and unobserved heterogeneity. Both Cobb-Douglas and translog functional forms are evaluated with multiple estimators, including standard fixed and random effects models, pooled and mean group common correlated effects (CCE) estimators, and random-trend models to account for shared upward trends among variables. The regression analysis examines the relationship between green patenting and key determinants such as R&D expenditure, human capital, and environmental policy indicators. The results consistently show a robust positive effect of domestic R&D, whereas the impacts of other factors exhibit greater variability. Methodologically, the findings highlight the sensitivity of coefficient estimates to unobserved heterogeneity and the choice of functional form.