A growing interest in R&D tax incentive policies has given rise to a large number of evaluations, which provide contrasting results about their effectiveness. Our meta- analysis aims to explain the heterogeneity found in the R&D tax incentive evaluations by the features of tax incentives. We document that on average R&D tax incentives stimulate R&D expenditures across two streams of empirical studies. However, this averaged effect is moderated by the underpinning features of tax incentives. Our samples evidence that the estimations linked to incremental bases and related to targeted rules towards SMEs drive the positive results found in the literature. Introducing a cap or a pre-approval process does not decrease the effectiveness of R&D tax incentives, allowing governments to monitor the indirect support needed to stimulate private R&D expenditures. Our results highlight the importance of setting up a clear and stable tax incentives framework. Sources of uncertainty regarding the timespan, the amount of the financial returns from tax claims but also the main criteria to apply are likely to decrease their effectiveness in the short run.