Repetition refers to the school practice of holding back, in a given grade, students that generally did not achieve the desired performance at the end of a school year. This practice has the goal to correct possible flaws in learning or adjust to situations of immaturity of the student. In the context of the debate about the advantages and disadvantages of this practice, this dissertation intends to verify the effects of failure on the student’s performance during the initial years of Elementary School. Those students took part in the Longitudinal Study of School Generation 2005 (Geres Project) from 2005 to 2008. Throughout the research the study aims to comprehend: [1] If the students that experience failure, even at the cost of an extra year of schooling, presented real gains of proficiency that justify the use of such measure; [2] in the case of real gains in learning, do this gains maintain during the initial years of the Elementary School? To try to answer these questions, a observational study was developed that compare two pupils groups, the ones who failed and the ones who got promoted, paired up through matching, and analyze the failure impact on the proficiency level in Reading and Mathematics from these groups of students in different moments of their school trajectory. Corroborating with several researches on the failure effects, the results show that the students who failed improve less than the students who were promoted along the years. When comparated in the same grade, regardless from the school grade that they have attended, the students who failed have better results, although the effect in Reading gets lost along the school grades.