Capital, Labour, Materials and Additional R&D Investment in Japan - The Issue of Double-Counting

R&D components are investigated in dynamic factor demand models using pooled Japanese data. Models without R&D, with R&D (double or) wrongly counted and (once or) correctly counted are compared by means of GMM estimates, (non-)nested GMM tests and residual analyses. The results indicate that R&D contributes significantly to the explanation of capital, labour, energy and materials demand in the Japanese manufacturing industry. Double-counted R&D is even preferred to not incorporating R&D as a separate production factor. After including R&D as a production factor and correcting for double-counting, there is however no unambiguous answer to the question which model utilizes the information of R&D best.