Based on my experience, the US mathematical model is based the dynamic of people getting into good cliques that then give you the support you need to establish a publishing record and tenure. This dynamic definitely does not reward mathematicians for going off an spending a lot of energy on a long-term project. Nor does it reward mathematicians who are inclined to generalize.
The US model for everything involves getting into good cliques...
The problem is that there isn't a viable alternative vetting system. A college degree no longer vets its holder as competent, so to determine ability, you have to either go to your trusted network or take a risk. Risks are by nature dangerous.