Whether females should prefer to mate with old males is controversial. Old males may sire offspring of low quality because of an aging germline, but their proven ability to reach an old age can also be an excellent indicator of superior genetic quality, especially in natural populations. These genetic effects are, however, hard to study in nature, because they are often confounded with direct benefits offered by old males to the female, such as experience and high territory quality. We, therefore, used naturally occurring extra-pair young to disentangle different aspects of male age on female fitness in a natural population of collared flycatchers because any difference between within- and extra-pair young within a nest should be caused by paternal genetic effects only. Based on 18 years of long-term data, we found that females paired with older males as social partners experienced an overall reproductive advantage. However, offspring sired by old males were of lower quality as compared to their extra-pair half-siblings, whereas the opposite was found in nests attended by young males. These results imply a negative genetic effect of old paternal age, given that extra-pair males are competitive middle-age males. Thus, offspring may benefit from being sired by young males but raised by old males, to maximize both genetic and direct effects. Our results show that direct and genetic benefits from pairing with old males may act in opposing directions and that the quality of the germline may deteriorate before other signs of senescence become obvious.