School socioeconomic segregation is regarded as a threat to education equity worldwide. A policy tool randomising school admission can be utilised to adjust school socioeconomic status (SES), and it has the potential to compensate students disadvantaged by family socioeconomic status. To test this claim, this study investigated a longitudinal data. The analytic models reveal the following main findings: (1) school SES is positively associated with students' academic performance; (2) the positive effect of school SES is consistent among students from different socioeconomic backgrounds; (3) overmatching to middle-SES schools academically benefits low-SES students; (4) the frog pond effect prevents low-SES and middle-SES students from obtaining advantages through studying in high-SES schools; (5) school factors are overwhelmed by family factors in predicting low-SES students' academic achievement. This research suggests that although students generally benefit from the rise of school SES, overmatching to a school does not necessarily compensate for students' academic disadvantages. Due to the potential frog pond effect, school socioeconomic desegregation fails to mitigate education inequity as a policy tool.