Nepotistic practices are detrimental for academia. An analysis of shared last names among academics was recently proposed to measure the diffusion of nepotism, the results of which have had a huge resonance. This method was thus proposed to orient the decisions of policy makers concerning cuts and funding. Because of the social relevance of this issue, the validity of this method must be assessed. Thus, we compared results from an analysis of Italian and United Kingdom academic last names, and of Italian last and given names. The results strongly suggest that the analysis of shared last names is not a measure of nepotism, as it is largely affected by social capital, professional networking and demographic effects, whose contribution is difficult to assess. Thus, the analysis of shared last names is not useful for guiding research policy.