We investigate short and long-term effects of early childhood education using variation created by a unique policy experiment in British Columbia, Canada. Our findings imply starting Kindergarten one year late substantially reduces the probability of repeating the third grade, and meaningfully...