Using household level data this paper provides systematic evidence on the employment
impact of the National Rural Employment Guarantee Scheme in three Indian states:
Rajasthan, Andhra Pradesh and Maharashtra. We model this as a two stage Heckman
procedure where we model selection for NREGS in the first phase and the determinants of
hours worked in the second. A number of significant insights into the employment impact of
the National Rural Employment Guarantee Scheme are obtained.