Up to now, large scale applied economic models with endogenous growth have represented innovation, as the engine of long term growth, eihter as a consequence of human capital, which does not completely endogenized innovation, or as fruit of investments in R&D. Unfortunately, this approaoch limits the full endogenization of innovation to sectors that invest in R&D that is to say mainly to manufacturing sectors. In this framework, sectors which not invest in R&D were not considered as innovative sectors by nature in economic modelling and their apparent productivity growth were assumed to come from spillovers from innovative sectors. Nevertheless, recent empirical studies question this restriction and an emerging literature enlarges the range of innovative activities (e.g. Carlo, Hulten and Sichel, 2005  and 2006 ). Another limitation of this approach comes from its unability to grasp the impact of the wide diffusion of Information and Communication Technologies (ICT thereafter) considered as an important source of performance improvement. This feature relies on the general purpose nature of these technologies (defined as General Purpose Technologies, GPT thereafter) inducing strong and large improvements in innovation capacities for users.
Therefore, in this context, a new theoretical framework has been developed for the innovation module of the NEMESIS model in order to take into account both the ICT investments as engine of growth and other innovative assets in addition to simple R&D investments.
The main purposes of these modifications consist (i) in considering innovative activities in a broader sense than pure R&D, by taking also into account expenditures in other intangibles assets, notably softwares and training, and (ii) in considering ICT as GPT, notably by including them in the innovation function in order to reflect their enabling feature. This is, at our knowledge, the first attempt of such improvement in a large scale macro-sectoral model1. The aim of this paper is to test the modifications of the innovation module that are reminded in the first part. In particular, it verifies whether these modifications question the previous policy assessments realised with the old version of NEMESIS or not. In addition, it tests whether this specification is in line with the theory and empirical observations, and highlights how such improvement enables to enrich the analysis of policies supporting innovations.