Whether artificial reefs installed in estuarine/marine waters function to produce more fish (enhancement) or simply to attract existing fish (attraction) is still under debate. Despite little resolution over this issue, artificial reefs are often considered for use as compensatory mitigation for damaged marine resources. We estimate the quantitative enhancement of fish production under 4 plausible scenarios: attraction, enhancement, enhancement with fishing, and attraction with fishing. Our intent is not to resolve the attraction-enhancement debate, but to quantify the uncertainty associated with using artificial reefs as compensatory mitigation. Pertinent parameters for production calculations (fish density by size class, length-frequency distributions, diets, behaviors, age-specific growth and mortality rates) were obtained from syntheses of findings from artificial reef studies conducted in coastal waters of the southeastern USA and from species life-history profiles. Year-round reef inhabitants were separated into 2 groups: those whose recruitment appears to be limited by available reef habitat (only 2 taxa) and those not augmented in recruitment but potentially enhanced in realized production by provision of refuges and reef-associated prey (15 taxa). Estimates of enhanced production in this latter group were discounted by an index of reef exclusivity in diet to give production credit in proportion to consumption of reef-associated prey. Estimates of annual production enhancement per 10 m2 of artificial reef ranged from 0 kg under the attraction scenario to 6.45 kg wet weight under the assumption of enhancement plus protection from fishing. Application of fishing reduced the enhancement estimate by 32% to 4.44 kg 10 m-2 yr-1. A 4th scenario of attraction with fishing may yield a net decline in production of a similar magnitude. In contrast to many natural structural habitats (seagrass meadows, oyster reefs, salt marshes, mangroves) that have dramatically decreased over past decades and are clearly important nursery grounds, evidence is weak that habitat provided by artificial reefs on the shallow continental shelf of the southeastern USA is currently limiting to fish production. Until convincing empirical evidence appears, high scientific uncertainty limits confidence in using artificial reefs as compensatory mitigation. Furthermore, even if augmented production were achieved, managing fishing impacts would be critical to achieving the expected production benefit.