Behavioural responses to reduce predation risk might cause demographic ‘costs of fear’. Costs differ among species, but a conceptual framework to understand this variation is lacking. We use a life-history framework to tie together diverse traits and life stages to better understand interspecific variation in responses and costs. We used natural and experimental variation in predation risk to test phenotypic responses and associated demographic costs for 10 songbird species. Responses such as increased parental attentiveness yielded reduced development time and created benefits such as reduced predation probability. Yet, responses to increased risk also created demographic costs by reducing offspring production in the absence of direct predation. This cost of fear varied widely across species, but predictably with the probability of repeat breeding. Use of a life-history framework can aid our understanding of potential demographic costs from predation, both from responses to perceived risk and from direct predation mortality.