Student turnover has many negative consequences for students and schools, and the high mobility rates of disadvantaged students may exacerbate inequality. Scholars have advised schools to reduce mobility by building and improving relationships with and among families, but such efforts are rarely tested rigorously. A cluster-randomized field experiment in 52 predominantly Hispanic elementary schools in San Antonio, TX, and Phoenix, AZ, tested whether student mobility in early elementary school was reduced through Families and Schools Together (FAST), an intervention that builds social capital among families, children, and schools. FAST failed to reduce mobility overall but substantially reduced the mobility of Black students, who were especially likely to change schools. Improved relationships among families help explain this finding.