According to the ‘masking hypothesis’, diploids gain an immediate fitness advantage over haploids because diploids, with two copies of every gene, are better able to survive the effects of deleterious recessive mutations. Masking in diploids is, however, a double-edged sword: it allows mutations to persist over tine. In contrast, deleterious mutations are revealed in haploid individuals and are more rapidly eliminated by selection, creating genetic associations that are favourable to haploidy. We model various mating schemes and show that assortative mating, selfing, and apomixis maintain the genetic associations that favour haploidy. These results suggest that a correlation should exist between mating system and ploidy level, with outcrossing favouring diploid life cycles and inbreeding or asexual reproduction favouring haploid life cycles. This prediction can be tested in groups, such as the Chlorophyta, with extensive variation both in life cycle and in reproductive system. Confirming or rejecting this prediction in natural populations would constitute the first empirical test of the masking hypothesis as a force shaping the evolution of life cycles.