This study examines whether participation in NREGA is influenced by the identity of locally-elected leaders or their political affiliation.
Background
NREGA, a rural workfare program, is one of India’s largest anti-poverty programs. The program guarantees 100 days of manual work to every household in the country, helping provide income support and employment to millions of low-income individuals across the country. However, local politics can negatively influence the implementation of large government programs such as NREGA– especially when local leaders have influence over who receives the benefits. For example, local leaders may decide to parcel out welfare benefits to their own communities rather than distributing benefits to those who need them most. Previous studies furnish evidence that political leaders in India tend to discriminate in favour of their own caste or community when implementing government schemes.
Approach
Researchers used secondary data to isolate the effect of a leader’s identity and political affiliation on the demographic composition of NREGA beneficiaries. They first collected detailed data on each individual participant from the Andhra Pradesh NREGA web portal. The author then aggregated this data at the Gram Panchayat and Mandal levels. The author merged this data with the Andhra Pradesh State Election Commission data to get the required information about the gender and caste of the candidates and the Mandal Parishad Territorial Constituency elections information such as the race, party affiliation of the candidate, etc.
Since the study uses secondary data, it is difficult to find the exact causal effect of a leader’s identity on NREGA outcomes, as voters usually consider a candidate’s caste, gender etc. before electing him. In other words, the leader’s identity is endogenous, and a simple analysis would simply produce a summary of the electorate’s preference. To analyze the impact of political affiliation, the study looks closely at districts where the winning party “Congress” contested a very close election (the assumption being that these districts would be similar otherwise).
Key Findings
The study found that reserving a sarpanch seat for a specific caste or gender does not have very significant impact on the population that participates in NREGA. For example, researchers find that reserving a sarpanch seat for a Backward Classes candidate only increases the percentage of Backward Class workers in the NREGA program by 2 percentage points. Similarly, reserving a sarpanch seat for a Scheduled Caste (SC) candidate increases the share of SC workers in the NREGA program by only 2.4 percentage points. Surprisingly, reserving a sarpanch seat for a Scheduled Tribe (STs) or a woman candidate does not significantly affect participation rates, wages or the demographic composition of workers participating in NREGA. Researchers find that the effect of party affiliation on NREGA outcomes is not very significant – the difference between NREGA outcomes of the areas where the Congress party won (Congress was elected in the State) and constituencies where Congress loss was very small. Researchers found little evidence to suggest that the caste, gender or local leadership significantly influenced who participated in NREGA. Policymakers should study the structure of Andhra Pradesh’s program to identify features of the program that may successfully prevent local leaders from distributing benefits to individuals who share the same background as them. (For example, unelected Mandal-level officers like engineers exercised influence over how program money was spent at the local level, a unique feature of AP’s program).
Implications
Implementing a large-scale government program transparently and effectively may help to prevent capture at the local level. Further research is required to check whether it is the nature of the NREGA program (work-for-cash) that makes it effective at guarding against local leader biases or whether Andhra Pradesh’s program was unique in its success at blocking discriminatory implementation.