Grade retention is a widely used educational policy promoting human capital. However, its benets and costs are still under debate. Retention may aect learning, cognitive and psychological capacities, educational attainment and the lifetime income (through the timing of entry to the labor market). This paper estimates the causal eects of grade retention on all these outcomes exploiting a retention rule based on a school-GPA cuto that enables a regression discontinuity design. I use 15 years of administrative data from the population of students in the educational system in Chile. The ndings are fourfold. First, (marginally) retained students improve their cognitive ability (measured by standardized test scores) in the short and the long-run. Second, induces parents to increase time investments and expectations. Third, enhances student psychological traits, and overall learning experience. Fourth, achieve the same amount of education than (marginally) promoted, high school enrollment and graduation, but retained enter the labor market 0.4 years later for two of the four grade levels analyzed.