This study was carried out in the context of the reform of the Common Agricultural Policy (CAP) for the period 2021-2027: one of the key elements of this reform is the shift towards an obligation of results for some subsidies. Supported by specific cases, we first show that the distinction between the obligation of means and the obligation of results is overly simplistic. The pure obligation of results in the environmental field never truly exists, and practical examples fall on a continuum of estimates of results with varying degrees of accuracy. An estimation of the costs of six instruments found on this continuum (Green Payments (GPs), Agri-Environment- Climate Measures (AECMs), organic conversion support, High Environmental Value certification (HEV), and two carbon certification systems) enables us to draw several conclusions. First, the obligation of results is not necessarily more costly than the obligation of means: AECMs for example, which are generally considered as obligations of means, are more expensive to administer than carbon certification systems, which are typically considered as obligations of results. The genericity of the instrument plays a key role, making it possible to spread the design and monitoring costs across a large number of farmers. Next, as regards the effectiveness of the instrument in terms of environmental impact, working towards an obligation of results does not appear to be decisive per se. Two factors are, however: the ambition of the instrument and the level of additionality required, for example by making subsidies conditional upon demonstrating an improvement over an initial state. Finally, the specific advantage of shifting towards an obligation of results seems to be that it facilitates the environmental assessment of the CAP, which would make it possible to redirect support where necessary according to this impact data, which is currently unavailable. The reform of the CAP opens up the possibility of introducing new types of payment in the context of the eco-schemes under the first pillar, and especially the carbon certification systems. Indeed, these systems give a good deal of attention to the issue of additionality. Since they are neither more costly to implement nor less effective than an AECM type instrument, they could begin to emerge within the CAP. Moreover, the example of support for organic agriculture shows that basing CAP subsidies on external labels is not without precedent.