I calibrate an eco-epidemiological age-structured Susceptible-Infected-Recovered (SIR) model of the B.1.1.7 covid variant on the eve of the vaccination campaign in France, under a stop-and-go lockdown policy. Three-quarters of the welfare benefit of the vaccine can be achieved with a speed of 100,000 full vaccination per day. A 1-week delay in the vaccination campaign raises the death toll by approximately 2500, and it reduces wealth by 8 billion euros. Because of the large heterogeneity of the rates of hospitalization and mortality across age classes, it is critically important for the number of lives saved and for the economy to vaccinate older people first. Any departure from this policy has a welfare cost. Prioritizing the allocation of vaccines to the most vulnerable people save 70,000 seniors, but it also increases the death toll of younger people by 14,000. Vaccine nationalism is modeled by assuming two identical Frances, one with a vaccine production capacity and the other without it. If the production country vaccinates its entire population before exporting to the other, the global death toll would be increased by 20 %. I also measure the welfare impact of the strong French anti-vax movement, and of the prohibition of an immunity passport.