A Novel Pathway from Quotas to Violence Reduction in India
In Diversity, Violence, and Recognition, we find that on average, recognizing ethnic groups as part of the body politic - whether symbolically or with group differentiated rights - leads to less political violence. Always interested in learning more about the effects of recognition and inclusion on peace and violent outcomes, we were excited to read new work by PhD candidates Drew Stommes (Yale) and Aidan Milliff (MIT). Their study (forthcoming at Journal of Peace Research) examines the effects of quotas for Scheduled Tribes in Chhattisgarh, India, an area with high-intensity insurgent activity. Like us, they find that the quotas reduce violence. But the mechanisms at work differ from symbolic inclusion, addressing mistrust, or improvements to marginalized group members’ wellbeing upon which we focus. Stommes and Milliff find that inclusion of marginalized groups creates a new channel through which ruling authorities and security forces learn about insurgents. Thank you, and welcome to our blog, Drew and Aidan!
Dilemmas in adoption of inclusive institutions
Inclusivity is vital for maintaining healthy democratic institutions. During civil conflict, however, fostering inclusion—and in some cases numerical over-representation—of civilians from marginalized ethnic groups that constitute the primary pool of support for rebels might be politically unpopular. Our study–“Descriptive Representation and Conflict Reduction: Evidence from India’s Maoist Rebellion”–assesses the impact of greater inclusion in democracy for historically-marginalized communities on levels of rebel violence. Specifically, we study mandated representation for Scheduled Tribes (STs) in India’s village councils (gram panchayats), asking whether this policy measure affected Maoist insurgent violence in the central Indian state of Chhattisgarh. We find that insurgent violence declined in localities where such reservations were implemented, and we identify a novel information-based mechanism that can explain this result.
Our study builds on extensive academic research studying the socioeconomic and political effects of institutionalized inclusion. For instance, a seminal study in economics identified that reservations for women in village councils in two Indian states increased public goods investment in infrastructure more directly relevant to women’s needs. More recent work on gender quotas in India adds some nuance. As Rachel Brulé argues elsewhere on this blog, gender quotas increase confidence in government and protect important economic rights like inheritance for women living in favorable “intra-family” power configurations, while women with limited intra-household bargaining power face backlash from male relatives. Outside the household, recent analysis also demonstrates that ST reservations in India improved the implementation of a large workfare program, generating greater employment for ST community members. Additionally, King and Samii (2020) identify how formal recognition for ethnic groups can reduce political violence and improve economic vitality.
We study a mandate from the Government of India requiring villages in geographically-delimited regions to reserve certain village council seats for members of ST groups. Our analyses evaluate whether this mandated inclusion affected insurgent violence in the state of Chhattisgarh, the heartland of a long-running insurgent conflict in India. In 1996, India’s central government enacted the Provisions of the Panchayats (Extension to the Scheduled Areas) Act (PESA) to improve local governance for tribal groups in subdistricts–roughly equivalent to a county in the U.S.–designated as “Scheduled Areas” (SAs). For all gram panchayats located inside SAs, PESA required that the council chair and half of non-chair positions must be filled by ST members. These representation mandates were a blunt instrument for fostering inclusion: in the regions of Chhattisgarh that we study, nearly 30% of villages that received mandates for majority-ST councils did not actually have majority-ST populations. A roughly equal proportion of villages without quotas had majority-ST populations left unprotected by a representation mandate.
Studying Inclusive Institutions in Contexts of Ongoing Violence
These reservations were implemented in Chhattisgarh’s SAs in 2005, amid the backdrop of a Maoist rebellion often referred to as the “Naxalite” insurgency. This Maoist rebellion began in the early 1980s within the present-day states of Chhattisgarh, Jharkhand, and Andhra Pradesh. Insurgent state-building has been especially concentrated in the rural forests of southern Chhattisgarh. This insurgency presented a major security challenge to the Indian state; one estimate suggested that the Naxalites once controlled 40,000 square kilometers (roughly the size of Switzerland), and were active in 1/3 of all districts in the country.
Chhattisgarh contains a disproportionately large ST population and numerous subdistricts designated as SAs. Myriad academic studies and journalistic accounts demonstrate how members of ST groups in Chhattisgarh constitute an important pool of civilian support for the Maoist insurgents, providing vital information about counterinsurgent operations and directly bolstering insurgents by joining the fight as combatants.
To study whether gram panchayat reservations for ST groups in Chhattisgarh’s SAs affected insurgent violence, we implemented a geographic regression discontinuity (RD) design. This empirical approach compares the average level of insurgent violence for villages just inside SA boundaries against violence in nearby villages located barely outside SA boundaries. Our primary analyses draw upon insurgent-perpetrated security force fatality data for Chhattisgarh geocoded at the village level. Spatial data for Chhattisgarh’s villages along with detailed lists from the Indian government identifying which subdistricts are designated as Scheduled Areas allow us to determine which settlements to consider in our geographic RD analysis.
Inclusion Reduces Violence
The main results from the geographic RD set-up indicate that treated villages (i.e., those with ST reservations mandated by PESA) experienced, on average, 0.07 fewer security force fatalities from 2005–2014 than villages outside Scheduled Areas. This village-level effect is equivalent to 15 fewer fatalities at the subdistrict-level and approximately 89 fewer fatalities at the district-level. The average subdistrict and district-level fatality counts after the 2005 elections reached 9.45 and 45.47, respectively. Considering these estimates along with a battery of robustness checks, we found compelling evidence indicating that reservations for STs in gram panchayats reduced insurgent violence against security forces. They also likely reduced violence against civilians given that civilian and security force fatalities are strongly correlated at the district level.
How Did Inclusion Reduce Violence?
These regression discontinuity estimates say little, however, about why reservations reduced rebel attacks. To this end, we examined whether ST reservations increased ST community members’ economic well-being and affinity towards the Indian state. King and Samii’s cross-national study of ethnic recognition shows that these are two mechanisms through which inclusion works. Inclusive institutions can help address tangible inequalities or even reduce the salience of ethnic identity as intergroup contact builds trust across communal lines. Their findings, mixed with the “hearts and minds” counterinsurgency philosophy of the early 21st century, yield a plausible explanation for the relationship between ST quotas and violence decline: If reservations improved individual-level economic outcomes and increased civilian support for the state, they may have reduced civilian support for the insurgency, undermining rebels’ capacity to perpetrate attacks. Using village level data on socioeconomic indicators and political participation, we were surprised to find little support for these explanations in the case of Chhattisgarh.
Instead, qualitative evidence from the conflict suggests a novel mechanism at work. Reservations brought ST community members closer to the state and security forces. This provided state and security forces with a windfall of valuable information about the rebellion. This information started to flow even though there was little evidence that ordinary ST members’ attitudes toward the state were changing. According to numerous reports, ST members occupying roles in gram panchayats regularly provided intelligence related to insurgent operations to police and were subsequently killed by rebels for being local informers. Panchayat members have routine, day-to-day interaction with their constituents, meaning that these elected officials are much more likely to accumulate intelligence regarding rebels which they can share with state security forces. According to Sundar (2018), insurgents in areas with predominantly ST populations grew deeply suspicious of individuals holding the position of council chair, given the close relations these local officials would often develop with arms of the Indian state. This information likely provided a tactical advantage to counterinsurgents who could then more effectively operate against rebels.
A good news story?
While we see this reduction in insurgent violence as a net good, we caution against drawing overly-optimistic conclusions from the success of ST reservations. First, our results indicate a difficult tradeoff: while ST reservations reduced rebel violence, ST community members getting involved in local politics faced a greater risk of being killed by rebels on the suspicion of being counterinsurgent collaborators. Additionally, we find limited evidence that ST reservations increased state legitimacy of ST members’ economic well-being. While the reduction of insurgent violence we identify could serve as a precursor for improving these outcomes in the long run, ST reservations in Chhattisgarh in the short- and medium-term did not generate significant improvements to residents’ well-being and increased local elected officials’ risk of death.
