R&D in Worldscan

Given the prominent role of innovation in economic policy, the analysis of innovation is also important for policy-oriented economic research. A CGE model is a choice vehicle for the analysis of the long-term economic consequences of innovation policy. So far, attempts of such an analysis have however been rather scarce. In addition, in models that do incorporate innovation, R&D is frequently modelled as an exogenous rather than an endogenous event. Exceptions are the models Quest III, Mesemet and International Futures. WoldScan, the CGE model for the world economy of CPB Netherlands Bureau for Economic Policy Analysis, includes endogenous R&D, thus providing a flexible instrument for evaluating the effect of innovation policy.

This memorandum analyses and checks whether the effect of R&D on total factor productivity (TFP) as modelled in WorldScan is consistent with the empirical estimates used for the calibration of the model.

We find that the model deviates from these empirical estimates in two ways: i) the effect of R&D in WorldScan differs considerably over countries and sectors. Some validation for this sectoral pattern is found in the empirical literature; and ii) in WorldScan, R&D not only generates spillovers, but is also treated as a production factor. There is thus a potential for double-counting the effects of private R&D. We find that this risk is only partly realized. In addition, we have considered the effectiveness of an R&D tax credit in WorldScan and find that such a tax credit raises both R&D expenditures and welfare. The degree of effectiveness of the tax credit in raising R&D expenditures is broadly consistent with the empirical literature: 1 euro extra R&D credit raises total R&D expenditures by an equal amount.