The paper examines the effectiveness of social audit as a tool to enhance accountability by measuring the impact of social audit on the implementation of National Rural Employment Guarantee Scheme, the flagship employment guarantee program of the Government of India, in the state of Andhra Pradesh, using a mix of quantitative and qualitative methods.

The main research questions addressed are: the impact of social audits on the size of the program and the payment process; are social audits results good indicators of the overall quality of program implementation; compare the performance of Karnataka, a neighbouring state, which has not taken up social audit, to Andhra Pradesh, in the overall implementation of the program; and the reasons behind the successful scale up of social audits in Andhra Pradesh. A difference of difference estimator was used to estimate the effect of social audit using the person-days of work generated and the proportion of timely payments in mandals (sub-district level) where social audit had been conducted and mandals where it had not been conducted in the years 2006-07 and 2007-08.

The results showed that there is a statistically significant improvement in the size of the program as measured by the mandays generated. There was no statistically significant improvement in the proportion of timely payments, which can be attributed to technical problems in scaling up the payment process. It was found that the qualitative reports provided useful inputs on the process related aspects (performance of functionaries, maintenance of muster rolls etc) that were missing from the quantitative performance reports. It was found that the program is not in a very stable position in Karnataka, given the fact that there has been a decrease in the size of the program in the current year, and a comparison with Andhra Pradesh would not be a fair. An important insight was that the social audit program generated a great deal of public support in Andhra Pradesh, as manifested by the huge turnouts in the sub-district level meetings, which resulted in political support cutting across party lines. Another critical strategy was co-opting the lower bureaucracy in the entire process, so that there were no major problems during roll-out.

The overall conclusion is that social audits are indeed an important tool in building social awareness which results in a greater demand for work which translates into increased size of the program. The process also exposed corruption in the implementation of the program and a total amount of Rs 20 million of program funds was recovered.

Our recommendation is that the Andhra Pradesh experiment with social audit can be replicated elsewhere in the country, provided that the learnings from its example are internalized, the program is launched in an incremental manner and political issues generated by the process are carefully handled. It is specifically recommended that the Government of India should finance a pilot social audit project in two districts in each state of the country, roughly modeled on the Andhra Pradesh example. The states could then do a comprehensive roll out across all districts based on the state-specific learning from the pilot projects.