Many coral reef fish form transient spawning aggregations at sites located a few to hundreds of kilometres from their normal residence areas. Reef fish spawning aggregations ("FSAs") are often heavily exploited, which make them targets for management with marine reserves. We used a per-recruit model to compare the long-term conservation (impacts on female spawning-stock biomass-per-recruit (SSBR) and female : male sex ratio, SR) and fisheries effects (impacts on yield-per-recruit, YPR) of spawning reserves vs. normal residence reserves for two data-poor populations from Seychelles with contrasting life history traits and sexual modes: the Siganus sutor population of the main granitic islands, which has a fast life history and is gonochoristic and the Epinephelus fuscoguttatus population of Farquhar Atoll, which has a slow life history and is protogynous. Overall, our results suggest that normal residence reserves are more effective at improving both the SSBR and YPR of S. sutor . In contrast, the protection of a substantial fraction of spawning sites is preferable for E. fuscoguttatus to ensure the reproductive output of this protogynous population through normalization of SR and maintenance of high SSBR. Neither spawning reserves nor normal residence reserves improved the YPR of E. fuscoguttatus . However, yields of E. fuscoguttatus may increase on the long term via recruitment subsidy if a substantial fraction of spawning sites is protected. This may occur only if the population was recruitment limited in the absence of reserves and increases in SSBR compensate for lost opportunities caused by the area closures. Sensitivity analyses revealed that the relative effects of spawning reserves and normal residence reserves relate more to the change in catchability occurring with FSA formation than to life history traits. Thus, normal residence reserves should be preferred over spawning reserves for S. sutor essentially because its catchability at spawning sites is low relative to many other aggregation-forming populations. S. sutor therefore suffers higher fishing mortality in normal residence areas than at FSA sites. Our study demonstrates that spawning reserves are not always the most effective tool for balancing conservation and exploitation objectives for FSA-forming populations, and that this measure should ideally be weighed against other management options.