The spate of Maoist attacks on security personnel in Chhattisgarh this week serves as a reminder that Moaist insurgency is the single biggest internal security threat faced by India. This column analyses the impact of MNREGA on Maoist violence and finds a spike in police-initiated attacks on Maoists following the implementation of the job guarantee scheme in 2006. This is possibly because MNREGA provides credibility to the government’s commitment to development, making the local population more willing to share information on Maoists.
A large part of the debate on the effectiveness of India’s Mahatma Gandhi National Rural Employment Guarantee Act (MNREGA) among policymakers and academics has focused on the economic impact of the programme. Questions around the effect of MNREGA on employment, agricultural wages and productivity, and so on are undoubtedly very important. However, given the size and ambitiousness of the programme, it is also important to understand its broader impact. An analysis of the cost effectiveness of MNREGA should include all benefits and drawbacks of the programme, even those that go beyond the narrow economic consequences.
Impact of MNREGA on Maoist violence
One important area in which one may expect MNREGA to have an important impact is national security and the intensity of Maoist violence, more specifically. While Maoist insurgents have been active since the late 1960s, the intensity of the conflict increased significantly after the Communist Party of India (Maoist) was formed from previously competing groups in 2004. The government has acknowledged that military strength alone has not been a successful strategy in reducing the intensity of the conflict. In 2006, the then Prime Minister Manmohan Singh spoke of the Maoists as the “single biggest internal security challenge ever faced by our country”.
The aim of the Maoist insurgency is to obtain a liberated zone in central India since they believe that the lower classes are being neglected by the Indian government in favour of elites. Civilians in highly Maoist-affected areas play a crucial role in the conflict by aiding both the police and Maoists through information provision and other forms of explicit and implicit support (Mukherji 2012). Since areas with high Maoist violence intensity also tend to be economically underdeveloped, one might expect MNREGA to have an especially important impact, both through providing actual economic benefits as well as through making the government’s commitment to economic development in those areas more credible for the local population. This means that MNREGA may improve the relationship between civilians and the Indian government in Maoist-affected areas and thereby change the nature of the conflict and the intensity of civilian support for both sides. If this is the case, then MNREGA may have important broader impacts on daily life in India that go beyond rural labour-market effects.
Our paper looks at the effect of MNREGA on Maoist violence in the early days of the scheme; we exploit the fact that the programme was rolled out in different implementation phases over time according to an algorithm that prioritised poorer districts (Khanna and Zimmermann 2014). This feature allows us to compare districts with similar poverty levels and other socioeconomic characteristics at a time when some districts had received MNREGA benefits a year earlier than the other districts. This similarity helps rule out that there are systematic differences between districts with and without access to MNREGA that could possibly drive the results.
Using data from the South-Asia Terrorism Portal (SATP),1 we code up the location, date and number of fatalities of all Maoist-related incidents between 2005 and 2008 in 17 major Indian states. SATP collects all this information from media reports, including details on who initiated the attacks. This allows us to compare the amount of violence in Phase 1 districts that received MNREGA to similar Phase 2 districts that did not. Our dataset contains 418 districts, of which 158 are predicted to be in Phase 1 by the algorithm
We find that Maoist violence increases in districts after the implementation of MNREGA, with the effects concentrated in the first few months after the start of the programme in 2006. This increase in violence is driven by more police-initiated attacks, and Maoists are the most affected group in terms of fatalities, injuries and arrests. However, there is also some evidence of more Maoist violence against civilians.
These empirical patterns are consistent with the idea that MNREGA makes civilians in Maoist-affected areas more willing to support the government by sharing information on insurgents with the police. This makes police and military forces more effective at tracking down insurgents and leads to more deaths, arrests and injuries among the Maoists. At the same time, Maoists are increasing attacks against civilians to punish local communities suspected of being police informers. A number of newspaper articles on Maoist incidents that we use in our analysis explicitly mention that Maoists leave leaflets in villages after attacks on civilians that brand the affected individuals as police informers. At the same time, there is little evidence of the Maoists directly sabotaging MNREGA projects in our analysis. There is also anecdotal and quantitative evidence of actual economic benefits being low in the early stages of MNREGA, for example due to implementation delays and low implementation quality (see for example, Niehaus and Sukhtankar 2013, Zimmermann 2014). This suggests that the information-sharing effects are less likely to be directly linked to actual economic rewards, but more to the promise of future benefits of the programme.
Despite the short time window of our analysis, there is reason to believe that the effects we find are of consequence. The former Indian Home Secretary Gopal K. Pillai said in 2010, for example, that the intelligence gathering system of the police had improved over the last couple of years, making police forces more successful at catching Maoists. These developments are also recognised by the insurgents who are accusing the government of turning the local population into police informers and of using surrendered Maoists as sources of information (see, for example, excerpts from press reports from 2007 here). Between 2006 and 2012, the Maoist insurgency has been losing ground in stronghold states like Jharkhand and Chhattisgarh, for example, and has been forced to move out of Andhra Pradesh (Mukherji 2012).
Implementation is key
Our results show that MNREGA has had a major impact on the national security situation in India. These effects should be taken into account when thinking about the cost effectiveness of the programme and potential reforms of the scheme. It is important to note that the continued support of civilians towards the government due to MNREGA over a longer time period is likely to depend crucially on the actual realised economic benefits that the programme delivers rather than a continued promise of some future rewards. Improvements in implementation quality are therefore an important factor. More broadly, the government may benefit substantially from combining military operations with local anti-poverty programmes that make the local population willing to support the government and to provide information on insurgents.
- A website managed by the Institute of Conflict Management, New Delhi.
- Hindustan Times (2006), ‘Naxalism Biggest Threat: PM’, 13 April 2006.
- Khanna, G and L Zimmermann (2014), ‘Guns and Butter? Fighting Violence with the Promise of Development’, Working Paper.
- Mukherji, N (2012), The Maoists in India: Tribals Under Siege, London: Pluto Press, United Kingdom.
- Niehaus, Paul and Sandip Sukhtankar (2013), “Corruption Dynamics: The Golden Goose Effect”, American Economic Journal: Economic Policy, 5(4): 230-269.
- Pillai, GK (2010), ‘Left-wing extremism in India’, Lecture delivered at IDSA, 10 March 2010.
- Zimmermann, L (2014), ‘Why Guarantee Employment? Evidence from a Large Indian Public-Works Program’, Working Paper.