Social insect workers are often capable of reproduction, but will not do so in the presence of a fertile queen. In large societies, queens are expected to produce a pheromone that honestly signals her dominance and/or fertility, to which workers respond by suppressing the development of their ovaries and by preventing other workers from reproducing (worker policing). However, what maintains the honesty of such queen pheromones is still under discussion. The explanation that an honest queen signal evolves simply because it serves the interest of all colony members does not seem to hold, since it is undermined by the fitness benefits of direct reproduction of workers at the individual level. A better explanation may be found in the idea that queen pheromones are difficult to produce for subordinate individuals, either because policing workers attack them, or because queen pheromones are intrinsically costly chemicals. Here, I discuss some of the arguments for and against these hypotheses and the evolutionary scenarios that each would lead to.