In 1997, the state government of Madhya Pradesh (MP), India, launched the Education Guarantee Scheme to provide access to primary schools in the state's rural areas with the help of community participation. The scheme led to the rapid set up of over 25 thousand schools in a span of 5 years - an average of over 2 schools per 1000 rural children in MP between the ages of 6 to 15 in 1997. Exposure of an individual to the scheme was jointly determined by her state of residence as well as her age at the time of the intervention. Using the 2005-2006 round of the National Family Health Survey, this paper combines the two sources of exogeneity to estimate the impact of this unusual policy experiment on educational outcomes...