Blog
Development programmes, security, and violence reduction

Evidence from an insurgency in India

Implementing development programmes in conflict-affected areas is crucial for conflict as well as poverty reduction. The big question is how do you carry out these programmes successfully? Are there specific conditions under which development policies are effective? What is their impact on violence?

So far, evidence on whether development policies can reduce violence is mixed. Iraq is an example of where they have been successful, but in the Philippines violence has increased — possibly due to insurgents sabotaging government programmes. Therefore, a sufficient level of security could be a precondition for development programmes to work in conflict areas. Indeed, evidence from Iraq and Afghanistan support this idea. To test the hypothesis further, we looked at the Indian state of Jammu and Kashmir (J&K) and examined the factors that have been related to reduced violence in the area.

The conflict in Jammu and Kashmir

The low-intensity conflict is rooted in the dispute between India and Pakistan over the territory of Kashmir, which began with the partition of the Indian subcontinent in 1947. The dispute has led to two open wars, in 1947 and 1965, and brought the two countries close to war on several occasions.

The current armed insurgency started in 1989 in the Kashmir Valley, spreading over time to other parts of the state. The insurgency has been kept alive by support from Pakistan in the form of arms and training, while the Indian army runs a counter-insurgency operation together with the police forces. Between 1998–2014, it is estimated that the insurgency resulted in over 25,000 deaths. Violence has however been decreasing since the early 2000s.

How did violence decrease?

So why has violence in J&K decreased? We used information collected from newspaper reports by the South Asian Terrorism Portal (SATP) on insurgent, civilian and security personnel casualties and the number of incidents involving explosives as indicators of violence. Visualizing the data of total casualties in a graph (see figure), it does appear there was a gradual shift from a ‘high-violence regime’ to a ‘low-violence regime’ — a pattern that is visible for all the violence indicators.

Trend in the total monthly casualties. Graph: Authors
Notes: Line of control (LoC): The de facto border between India and Pakistan. The JaisheMohammed (JeM): A group that has carried out terrorist attacks in the region.
Source: Author’s illustration

We wanted to confirm whether this was the case and used non-linear time-series techniques to precisely detect structural breaks in the violence data. Using these techniques, we found that there has indeed been two ‘regimes’ of violence: a high-violence regime between 1998–2003 and a low-violence regime between 2007–14. The period in between, 2003–06, can be seen as a transition period between the regimes, during which a ‘smoothed’ structural break took place. The methods we used allow ‘the data speak for itself’ to inform us when critical changes in violence took place instead of testing whether certain events are associated with breaks in the time series.

Knowing when structural breaks have occurred is important as they often occur as a consequence of changes in policy. Having identified 2003–06 as the crucial period during which there was a shift in regimes, we then asked the question: what policies were implemented during this period?

What happened in Kashmir between 2003 and 2006?

During this period several security and development-related events took place:

  • In 2003, at the beginning of the transition period, India and Pakistan agreed to a cease-fire along their border and restored diplomatic dialogue (suspended since the 1999 Kargil war).
  • The construction of a 550-kilometer-long border fence between J&K and Pakistan was completed in September 2004. Indian security forces estimate that it has been successful in reducing the infiltration of insurgents from Pakistan.
  • The Prime Minister's Reconstruction Plan for J&K (PMRP) began in March 2005. Implemented by the then Prime Minister of India, Manmohan Singh, its objective was the long-term development of the state-building infrastructure, providing basic services and creating jobs. As of March 2014, over US$13 billion (INR780 billion) had been disbursed.
  • The roll-out of the National Rural Employment Guarantee Scheme (NREGS) to J&K started in 2006, at the later stage of the transition period. Being one of the largest safety net programmes in the world, NREGS guarantees 100 days of manual work at the minimum wage to all rural Indian households.
Were these policies effective in reducing violence?

Having identified structural breaks and the events which coincide with the breaks, we tested if these events were effective in reducing violence using district-level violence data.

Rural road in Pakistan. Photo: Curt Carnemark / World BankWe found that security improved at the border after the completion of the fence. While there were more insurgent casualties in border districts, civilian casualties in the area decreased after the completion of the fence. This suggests that security forces were perhaps better able to target insurgents, thereby reducing civilian deaths. The NREGS and PMPR programmes were also effective in reducing insurgent and security force casualties.

These results give support to our idea that there needs to be a certain level of security before development programmes can work effectively. In the case of J&K, the construction of the fence improved the security environment, which allowed large-scale development programmes to be implemented. The programmes then further reduced violence, particularly that directed towards civilians.

The views expressed in this piece are those of the author(s), and do not necessarily reflect the views of the Institute or the United Nations University, nor the programme/project donors.