Discussions on the use of marine reserves (no-take zones) and, more generally, spatial management of fisheries are, for the most part, devoid of analyses that consider the ecological and economic effects simultaneously. To fill this gap, we develop a two-patch ecological-economic model to investigate the effects of spatial management on fishery profits. Because the fishery effects of spatial management depend critically on the nature of the ecological connectivity, our model includes both juvenile and adult movement, with density dependence in settlement differentiating the two types of dispersal. Rather than imposing a reserve on our system and measuring its effect on profits, we ask: "When does setting catch levels to maximize system-wide profits imply that a reserve should be created?" Closing areas to fishing is an economically optimal solution when the value derived from spillover from the reserve outweighs the value of fishing in the patch. The condition, while simple to state in summary form, is complex to interpret because it depends on the settlement success of the dispersing organisms, the nature of the costs of the fishing, the economic and ecological heterogeneity of the system, the discount rate, and growth characteristics of the fish population. The condition is more likely to be satisfied when the closed area is a net exporter of biomass and has higher costs of fishing, and for fish populations with density-independent settlement ("adult movement") than with density-dependent settlement ("larval dispersal"). Rather surprisingly, there are circumstances whereby closing low biological productivity areas, and even sometimes low cost areas to fish, can result in greater fishing profits than when both areas are open to fishing.