High-stakes exams are often administered at designated test centers, requiring many students to test in unfamiliar environments. We investigate whether such arrangements impact students' test performance and, by extension, access to educational opportunities. Using unique administrative data from China's national college entrance examination between 2016 and 2018 and its random assignment of test centers, we find that students assigned to a non-home school score 0.14 standard deviations lower than classmates testing at their home school, and they are 3.8 percentage points less likely to be admitted to college. The performance penalty is most pronounced in STEM subjects and partly driven by longer travel distances. Furthermore, it has significant inequality implications: the penalty is especially severe for low-achieving students and those from disadvantaged backgrounds. As test centers are predominantly located in high-performing schools, such ostensibly neutral assignment policies may unintentionally exacerbate existing achievement gaps between privileged and less privileged groups. A back-of-the-envelope calculation suggests that exam location accounts for over 7.6% of the observed performance gap between students from test-center and non-test-center schools.