In this paper we analyze the efficiency of voluntary incentive-based land-use policies for biodiversity conservation. Two factors combine to make it difficult to achieve an efficient result. First, the spatial pattern of habitat across multiple landowners is important for determining biodiversity conservation results. Second, the willingness of private landowners to accept a payment in exchange for enrolling in a conservation program is private information. Therefore, a conservation agency cannot easily control the spatial pattern of voluntary enrollment in conservation programs. We begin by showing how the distribution of a landowner’s willingness to accept a conservation payment can be derived from a parcel-scale land-use change model. Next we combine the econometric land-use model with spatial data and ecological models to simulate the effects of various conservation program designs on biodiversity conservation outcomes. We compare these results to an estimate of the efficiency frontier that maximizes biodiversity conservation at each level of cost. The frontier mimics the regulator’s solution to the biodiversity conservation problem when she has perfect information on landowner willingness-to-accept. Results indicate that there are substantial differences in biodiversity conservation scores generated by the incentive-based policies and efficient solutions. The performance of incentive-based policies is particularly poor at low levels of the conservation budget where spatial fragmentation of conserved parcels is a large concern. Performance can be improved by encouraging agglomeration of conserved habitat and by incorporating basic biological information, such as that on rare habitats, into the selection criteria.