Using National Sample Survey data for rural India we examine the incidence of capture in two workfare programs in rural India: the Rural Public Works and the Food for Work Programs for 1993-1994 and 2004-2005 respectively. We discover a high degree of program capture among the general population. Among the traditionally backward groups in Indian rural society - but with considerable variation in their living standards - there appears to be a higher degree of capture among SC (Scheduled Castes), than among ST (Scheduled Tribes). Targeting among SC worsened over time. There was an increase in capture by the fourth quintile (of household per capita expenditure) of SC, ST and landowners. This may be reflective of a varying degree of collusion between the elite and the program implementing agencies (e.g. village councils) over time. Thus, potential benefits of workfare get undermined. We also provide evidence to suggest that income-based targeting could outperform social group based targeting.