Abstract There has been a strong push within natural resource management to incorporate spatial structure into management regimes. However, discussions surrounding the appropriate designs of spatial management have largely been conceptual. This paper develops a spatial econometric model of fishing location choice using non-confidential data from the Great Barrier Reef coral trout commercial fishery. Harvest location decisions are modeled as a function of spatial patterns of expected economic returns. The preferred spatially dependent econometric model is shown to outperform ordinary least squares and fixed effects models in out-of-sample forecasting. Estimates from the spatial model reveal spatial spillover effects in fleet harvest location behavior. In particular, harvest activity at any given site is equally sensitive to same-site economic returns and surrounding-site economic returns. The econometric results are illustrated using a fee-based policy simulation. Results suggest non-spatial management is characterized by two inefficiencies. First, heterogeneity between sites is averaged, resulting a fee that is too high or too low across space. Second, fees that are too high or too low affect the fishing effort in nearby locations.