Although the relationships between family income and student identification for gifted programming are well documented, less is known about how school and district wealth are related to student identification. To examine the effects of institutional and individual poverty on student identification, we conducted a series of three-level regression models. Students of poverty are generally less likely to be identified for gifted services, even after controlling for prior math and reading achievement. Further, school poverty predicts the percentage of gifted students identified in a school. Within-districts, even after controlling for reading and math scores, the poorer schools in a district have lower identification rates. Whereas students of poverty are generally less likely to be identified for gifted services, poor students in poor schools are even less likely to be identified as gifted. [This article was published in "Gifted Child Quarterly" v62 n1 p6-24 Jan 2018 (EJ1163395).]