Abstract : The aim of that thesis is to bring a micro level contribution to the economics of innovation when using original datasets on the French case. Here the manuscript is structured in three chapters: the labour market of scientists and engineers, the incentives for innovation in the private sector and the existence of spatial spillovers in the school to work transition for graduates in science and engineering (S & I). In the first chapter of this thesis, we focus on the labour market for scientists and engineers. First, we identify the different characteristics of this labour market. We then analyze the determinants of access to R & D for young graduates in S & I. Finally, we question more broadly on the valuation of R & D in private companies. We determine whether it pays, or not, for an engineer to track his career in R & D. This analysis of the labor market of scientists in industry shows that there is no direct return in R & D activities. This result raises the question about incentives. The second chapter is centered on those incentives for innovation in the private sector. The measurement for innovation is that the engineer has filed a patent, usually chosen indicator to measure innovation at the firm level. We are therefore interested in the variety of monetary incentives received by the inventors in private sector. The literature shows two types of monetary incentives: i) incentives related to performance of the inventor (wage level) ii) related to business performance (stock-options grants). We pay particular attention to inter-firm mobility of inventors. In the final chapter of this dissertation, in line with the spatial analysis of innovation in literature, we analyze the existence of spatial spillovers effect in the recruitment of graduates in S&I by companies for their R&D activities. We also seek to determine whether the human capital of scientists is more highly valued in the inner circle of the more efficient scientific clusters. Our hypothesis tests the consistence for extent putting into question the theory of spatial job search. Finally, we pay special attention to the return to scientific human capital within the competitiveness poles established by the French government in year 2005 modeled on North American scientific clusters.