Decentralization leads to functional and financial devolution of authority, induces more transparency in the system and puts emphasis on local needs. It creates platforms to voice and institutionalize the interests of various groups, and comes nearer to the public by making all tiers of government accountable directly to the people. However, several complexities govern this effort in India. First, implementation involves several official agencies, creating administrative gaps, lack of coordination and fiscal complexities. Second, Indian society is largely agrarian and rural marked by divisions of religion, caste and economic class. Third, the socially vulnerable and poor are often trapped in interlocking economic transactions with affluent landlords making it difficult for them to go for collective change. This research critically evaluates these dynamics taking the case example of the Mahatma Gandhi Rural Employment Guarantee Scheme (MGREGS). We followed a threefold methodology in order to understand the policy dynamics, namely, stakeholders’ interviews, primary survey data analysis and secondary data analysis. Case studies were useful in understanding policy implementation dynamics through field observations. The multi-level regression results reflected co-existence of agriculture and MGREGS dynamics, with their own rules and objectives, apparently contradictory, but leading to social and economic equilibrium. Searching explanations for these contradictory results led us to two larger understandings: 1) Actors are seen to take rational decisions based on local socio-politico-economic understandings of the world. 2) These decisions are, however, at multiple levels and at multiple action spaces. Institutional rational choice perspective proposed by Ostrom (1999) corroborated these findings. Latent variable path analysis results indicated the dynamics of civil society and administrative bureaucracy as a negotiated equilibrium that has the potential to transform governance. The process of institutional evolution was through endogenous institutional change process as proposed by Grief and Laitin (2004). Furthermore, the effect of socio-structural factors on institutional structures was observed. The results indicated two major understandings: 1) the formation and effectiveness of institutional structures were dependent on social capital and in turn on social structures and networks 2) the influence of formal institutional structures on local governance and social dynamics affect the governance network formation (Hertting, 2001). Scenario-wise regression analyses results revealed that the effort to improve social capital of the groups through governance procedures need to be mediated through ‘local’ social structures. Agent based model results indicated the following: 1) regulatory dynamics need not necessarily follow the trend of socio-economic dynamics. Instead, they were in line with advocacy dynamics (Sabatier, 1988), which in turn depend on the social structures and networks. 2) regulatory strategies were endogenous institutional rational choices, given the existing socio-economic structures and networks of the society. Hence, Institutional theories were observed to be instrumental in understanding the policy implementation dynamics in democratic decentralization setup.