People’s participation and social accountability are increasingly perceived as mechanisms that foster transparency, and offer a potential to improve public programme delivery in India. This was highlighted, most recently, in the discourse on school mid-day meals following the tragedy in Bihar.1 In the public work projects implemented under the Mahatma Gandhi National Rural Employment Guarantee Act (MNREGA 2005), ‘mandatory’ social audits prescribed by the Act intend to empower beneficiaries to scrutinise programme expenditures and to monitor and keep track of programme delivery. Despite widespread acclaims that social audits are low-cost and powerful participatory tools whose remit may successfully be extended to other public programmes, a key question is whether community monitoring has lived up to expectations. Specifically, has community monitoring helped reduce corruption and improve MNREGA delivery? We undertake the first attempt to rigorously assess the impacts of India’s only large-scale community monitoring initiative, in Andhra Pradesh (AP) (Afridi and Iversen, forthcoming).
The social audit process in Andhra Pradesh
AP is widely perceived as a model state in MNREGA implementation. A foremost reason for this stature is AP’s unique record of conducting regular and systematic social audits of MNREGA projects. In 2006, the responsibility for implementing these audits was entrusted to an autonomous arm of the Department of Rural Development (named the Society for Social Audit, Accountability and Transparency (SSAAT) since May 2009). This makes AP unique and distinct from other states of India where audits are either not conducted or implemented in an ad-hoc and unsystematic manner. Regular and standardised audits were carried out in all 23 districts of the state, with an average of over two rounds of audits per Gram Panchayat (GP) between 2006 and 2010.
The first step in conducting the social audit comes through a notification to the relevant sub-district (mandal) office requesting unrestricted access to worksite logs (muster rolls)2 and other relevant MNREGA project documents, with reference to the Right to Information (RTI) obligations.3 Next, a team comprising state and district auditors visits the mandal headquarter and recruits ‘village social auditors’ (VSAs). In a two-day intense workshop VSAs are trained in MNREGA rights and regulations, conducting social audits, and obtaining information under the RTI Act. The VSAs are MNREGA beneficiaries and residents of the mandal.
The social audit teams then, over a period of about one week, organise social audits in all GPs within the mandal. In each GP, official labour expenses are verified by visiting labourers listed in the muster-rolls. Complaints by individuals or groups of beneficiaries and by the audit team are recorded and attested using a standardised audit report template. For verification of material expenditures, the audit team is required to inspect worksites. Except for the more obvious and easy-to-detect irregularities such as ‘ghost’ or non-existent projects, the verification of material expenditure is typically perceived as more complex when compared to labour expenses.
Once audits of all GPs are completed, a mandal-level ‘public hearing’ to discuss the audit findings is organised with mandatory attendance for all officials implementing MNREGA projects. Attendees typically include ‘‘wage seekers from the villages in the mandal, the social audit team, branch post-master, key implementing officials, members of the vigilance cell, elected representatives and a district-level ombudsman’’ (Aiyar et al. 2013). Complaints are read out, testimonies verified, and accused officials are given an opportunity to defend themselves.
After the ‘public hearing’, a decision taken report (DTR) is created by the officer presiding over the hearing. In this report, the responsibility for each confirmed malfeasance is pinned on a single or multiple programme functionaries.
Thus, the social audit process in AP uniquely combines a top-down approach (SSAAT controls the timing and overall conduct of audits) with grass-roots participation (involvement of VSAs and MNREGA beneficiaries).
Analysing the impact of social audits
Conceptually, even if a state government was to publicly announce regular social audits, the first round, because of limited state credibility, is likely to take public officials and transgressors by surprise. To start with, the cost of effort for transgressions in basic programme delivery is likely to be low. Further, local MNREGA beneficiaries are expected to have high stakes in employment availability and in timely and due pay, as well as sufficient initial capacity to detect transgressions in basic programme delivery. We anticipate an evolving dynamic process with more effective local participation through learning and improved auditing with repeated audits. The drawback is that transgressions may also become more sophisticated. Thus, monitoring may result in substitution of one type of irregularity for another as transgressors learn to manipulate the new system and discover other avenues for rent extraction (Olken and Pande 2011). Hence, while easy-to-detect irregularities are expected to decline with repeated audits, hard-to-detect irregularities may surge in later audit rounds.4
Testing these predictions is made harder both by the fact that social audits were not rolled out randomly in AP and by the type of information (complaints) recorded by social audit teams.
In our analysis, we use a panel data set assembled through meticulous extraction and translation of information from official social audit reports from 2006 to 2010. This panel contains data from up to three rounds of social audits from close to 300 GPs in eight AP districts. We focus on whether programme performance, measured by irregularities in programme implementation are affected by additional audits within the same mandal. Thus, we prioritise outcomes that relate directly to malpractices and irregularities that speak to widespread concerns about leakages and corruption in large public programmes in India. Our analysis accounts for other factors that could impact corruption and the quality of programme delivery: (i) changing beneficiary awareness and growing confidence in the integrity of the audit process, and (ii) improved audit quality as audit team members become better at identifying discrepancies.
We find that the aggregate number of irregularities were higher in audit 2 relative to audit 1, whereas there was no significant difference between audits 2 and 3. While the aggregate number of labour-related irregularities did not change between audit rounds, there was a secular rise in irregularities pertaining to material expenditures in MNREGA projects that are, on average, harder-to-detect than those relating to the labour component.
Disaggregating, the labour-related irregularities we find that there were more complaints about non-payment or delay in wage payments in audit 2 relative to audit 1 but the increase in such complaints was insignificant over the three audit rounds. Interestingly, the number of complaints related to non-provision of MNREGA work declined significantly in round 2 vis-à-vis round 1. But there was no effect of audits on reducing the number of harder-to-detect labour irregularities (‘ghost’ workers, bribes etc.) reflecting corrupt practices.
The increase in materials-related complaints in audit 2 was driven by the rise in ‘ghost’ projects, bribes (including inflated bills) and non-provision of official records on materials expenditure. However, by audit 3, the number of non-existent projects was not higher than in audit 1. Thus, the increase in materials-related complaints in audit 3 was driven by the significant increase in bribes/ inflated bills and non-provision of expenditure records.
The impact of audits on other programme outcomes – employment generation, targeting of Scheduled Caste/ Scheduled Tribe (SC/ ST) population - was also insignificant.
Despite beneficiary learning and the greater capacity of the participatory audit process to detect irregularities in later audit rounds, the overall social audit effects on reducing easy-to-detect malpractices was mostly absent. This failure, together with the overall rise in hard-to-detect irregularities, signifies a change in the anatomy of corruption and a failure of the social audit process to effectively deter programme malpractices and leakages.5
So what explains the absence of an effect of social audits on reducing programme irregularities? Analysis of administrative data on social audit findings in AP suggests that during the period of our study, follow-up and enforcement of punishments was weak in the study districts. While this may have been mitigated by the establishment of vigilance cells in AP post 2010, less than 1% of irregularities where one or multiple programme functionaries were held responsible led to termination/ removal from service or criminal action. Even modest punishments such as suspensions, show-cause notices or being deemed ineligible for contractual work were meted out for less than 3% of these irregularities: 87% of the missing amounts were yet to be recovered as of 2013.
It is noteworthy that responsibility for MNREGA implementation is held collectively by grassroots and block-level functionaries. Yet, a majority of the irregularities upheld during public hearings was pinned on a single GP-level, contractual functionary – the field assistant – who is typically a GP resident. The ‘naming and shaming’ during the public hearing, might have been an effective deterrent for this particular functionary. However, such social sanctions are unlikely to curtail malpractice among other programme functionaries who often escape responsibility (higher level functionaries like block development officers or assistant programme officers, and/ or non-residents of the GP such as branch post masters).
Hence, a key lesson from our analysis is to ensure that social audits culminate in the type of enforceable and credible ‘contract’ that allocates responsibilities, defines timelines for punishments, and ensures that those who found guilty of irregularities are promptly penalised. The credibility of the social audit process rests ultimately on the ability and willingness of the political and bureaucratic establishment to take effective remedial action against offenders.
A second, critical take away is that without sufficient institutional support, the expectation that beneficiary-led audits would spontaneously arise is unsustainable. Systematic and regular audits with beneficiary participation have not taken off in other parts of the country despite the Act mandating them and laudable efforts by civil society organisations. Thus, to ensure effective bottom-up involvement, beneficiary participation must be induced and strengthened through a combination of top-down and grass-roots approaches.
Third, the capacity- and experience-based skill accumulation by core social audit personnel must be bolstered in order to address the dynamic shifts in the nature of malpractices. Technical expertise required for auditing materials expenditures may be critical for addressing the increase in second-generation programme leakages.
Finally, our work underlines the need for incorporating rigorous programme evaluation in the roll out of social audits and obtaining objective measures of program leakage. Greater vigilance in the documentation of social audit evidence, which could be critical for assessing the impact of these audits, is also essential. Careful documentation of audit findings would require improvements in the training of auditors and greater emphasis on the importance of more complete documentation during the course of the training modules.
To conclude, in 2011-2012, MNREGA had provided employment to almost 40 million households across the country at an annual expenditure of almost Rs. 40,000 crores. The cost of social audits amounts to a mere 1% of this expenditure. Hence, while the potential benefits of public programmes are large, the costs of ensuring that those benefits are realised through beneficiary led audits are low.
- Muster rolls maintain the list of labourers, daily attendance etc. at each worksite
- The Right to Information Act (RTI) 2005 mandates timely response to citizen requests for government information.
- Easy-to-detect irregularities include non-payment of wages and ghost projects. Hard-to-detect irregularities include ghost workers, bribes, quality of materials and absence of expenditure records.This hypothesis assumes that audits effectively detect malpractices and carry a credible threat of punishment for malfeasance.
- These findings are robust to a variety of checks including restricting the data to irregularities registered only by the audit team which, in principle (and unlike for household or individual complainants) should be immune to any biased reporting of programme malfeasance, and the intensity of the audits at the mandal level.
- Afridi, F. and V. Iversen (2013), ‘Social Audits and MGNREGA delivery: Lessons from Andhra Pradesh’, forthcoming in Barry Bosworth, Arvind Panagariya and Shekhar Shah (eds.), Brookings-NCAER, India Policy Forum (Vol. 2013), forthcoming.
- Aiyar, Y., S. K. Mehta and S. Samji (2013), ‘India: Implementing Social Audits’, in K. Subarrao et al. (ed.), Public Works as a Safety Net: Design, Evidence and Implementation, Washington DC: The World Bank, Chapter 11, pp. 249-68.
- Olken, B. and R. Pande (2011), ‘Corruption in Developing Countries,’ NBER, Working paper no. 17398.