Résumé : This paper provides an in-depth and critical analysis of the current Belgian fiscal regime towards business research and development (R&D) expenditures. The paper clearly shows that what matters the most is the way a policy decision is implemented. In this respect the current R&D tax credit scheme is subject to several drawbacks. The first one is the very low level of support for ‘additional’ researchers. The second one is the ‘nominative’ aspect of each tax credit: the deduction is allowed for well-defined employees. The third main drawback is the non-permanent characteristic of the tax credit. Once a deduction is allowed for an employee, it must be validated for each subsequent year. These drawbacks are associated with a high level of complexity and a heavy administrative burden, for a weak or zero effectiveness of the scheme. We suggest several avenues for improvement, including a drastic simplification of the policy design and a much higher generosity, especially if the Lisbon objective is to be reached.