The hypothesis that the history of management compliance (strong versus variable), closure size (0.3 to 28km2), closure age (2 to 39 yr), and habitat/geographic variables influenced total fish biomass, the number of species, and the ratio of herbivores to carnivores in no-take areas was tested. Twenty existing closures across 6 countries in the western Indian Ocean were sampled during the same time period (2004 to 2007). Geographic location and closure attributes were weak to moderate predictors of these coral reef fish community variables for analyses of all closures. Strong interactions between fish community variables and levels of compliance indicated that high compliance sites largely drove the closure-fish community relationships. Variable or weak compliance and closures <1km2 exhibited limited recovery of fish communities. The closure area-fish biomass data for strong-compliance closures indicates that biomass is stable above 5km2. A mix of community-based, private, and national program closures were sampled, and, although low sample sizes within each of these management systems precludes conclusive analysis, none of these management systems was universally successful in terms of compliance or maximizing fish community variables.