In most countries, governments intervene in the process of R&D by financing a substantial part of it. The mechanism employed for choosing the projects to be financed is a committee composed of experts who evaluate projects in their field of specialisation, and decide which ones should be funded. This method is conservative. Proposals for new ideas are too often rejected, and inventions are commonly thrown out of the set of potential projects. In this paper, I propose a mechanism that will allow less conformity: focal randomisation. This states that projects that are unanimously ranked at the top by all reviewers will be adopted. Projects perceived as valueless by all are rejected, while projects that are ranked differently are randomised. I compare the average return under the present and proposed mechanisms. I examine under which conditions this new method is preferable, and its consequences on economic growth. N MOST COUNTRIES, governments intervene in the process of R&D by financing a substantial part of it. The reason for this intervention is that research and development undertaken by one firm has positive spillover effects on the entire economy. Since firms do not take these spillover effects into account, they invest in R&D less than the optimal amount. It is difficult to estimate these externalities, but it could double the real rate of return. Moreover, basic research is a non-excludable investment and firms have no interest in undertaking it in a competitive market. Therefore, government financing is necessary. In some countries, it can amount to more than 60% of total R&D funding. Such an intervention, however, creates a problem: how does the government choose which projects to finance? It could be that the projects chosen are not those with the highest potential growth for the economy and therefore not optimal for the country. In this paper, I discuss the mechanism of evaluating and choosing the projects to be financed in a given field, but do not treat the problem of deciding which field to finance. The mechanism employed in most countries is a committee composed of research fellows who evaluate projects in their field of specialisation and decide which ones should be funded. Alternatively, the decision is left to the committee chair, taking into consideration the referees’ reports. Both mechanisms are based on what is termed ‘peer review’. The problems with peer review have been analyzed at length and, in the next section, we give an outline of the different problems. However, most of these negative effects do not affect the rate of growth of the economy as does one specific problem: the conservative bias. Indeed, peer review presents a bias against innovative applications, as emphasised by many and especially Martin (1997: 72):