Much of the motivation to develop effective programs of biological pest control is derived from the desire to reduce the negative impacts of agricultural pesticides on human health and the environment. What are the major drivers of pesticide use? We address this question with a broad comparative analysis of pesticide use, including pesticides targeting arthropods, plant pathogens, and weeds, across the main crops grown in California. We find that there are both economic and ecological drivers. As predicted by agricultural economics theory that models farmers as profit maximizers, the main driver of pesticide use is crop value: crop value explains much of the variation in pesticides targeting arthropods and plant pathogens. Crop value does not, however, explain variation in herbicide use, where there are abundant and effective non-chemical control options. Ecological drivers are also important: crops whose marketed parts are belowground receive fewer pesticide applications targeting arthropods, and herbicide use is greater on perennial crop plants than on annuals. We also find that use of pesticides targeting arthropods and plant pathogens is lower for crop plants that are phylogenetically more isolated from native California plants. This suggests that host-switching events are important in the assembly of pest communities, and that switching is difficult when crop plants are only distantly related to native plants. We examine variation across crops in the source of crop pests (native versus introduced) and their diet breadth to paint a picture of crop pest community assembly that forms the context for biological control programs.