VoxEU Column Frontiers of economic research Health Economics Migration

Capturing ungoverned space and the safety of children

The bloody conflicts in Syria and Iraq have forced the issue of refugees onto the global agenda. However, among the neglected aspects of this discussion are how governance can be restored to conflict regions and the welfare effects that such actions, which are likely to be coercive, will have on local residents. This column examines the impact of a counter-insurgency programme in the Philippines on one development outcome in contested territories – malnutrition of young children. The programme saw a substantial long-term decrease in malnutrition in recaptured areas, but a rise in malnutrition in neighbouring areas. Such efforts may simply displace insurgents and their negative effects, rather than reducing them.

The atrocities in Syria and Iraq committed by Daesh (often called ISIS), together with the collapse of government control in parts of those two countries, have created the largest flow of refugees in recent history, and ignited an important public discussion about ungoverned spaces. Europeans would of course prefer that stable governance be somehow returned to Syria and Iraq, as well as to the other sources of refugee flows – Somalia, Afghanistan, Pakistan, Yemen, Libya, and other countries in which the risks of remaining in ungoverned spaces outweigh the perils of fleeing.

How could governance return? Rosy euphemisms aside, the most likely answer is that the existing government consolidates its control of territory through coercion. If Europeans support those governments in consolidating control over ungoverned spaces, the next question must then be: What effect would that coercive act have on local residents?

The answer may be obvious for Syria, Iraq, and IS, judging by refugee flows, but it is not obvious in general. While we traditionally believe that once consolidated, states provide important public goods and secure property rights, many insurgent groups provide services as well (Keister 2011, Heger and Jung 2015). Moreover, since the expansion of state authority is likely to be violent, improvements in institutions might well be offset by the risks and human rights abuses accompanying a transition.

Rebellion and counter-insurgency in the Philippines

The Philippines provides a setting to address this question. Its government has for decades attempted to consolidate control over peripheral areas. Active rebel groups include Marxists, Islamist separatists and the Abu Sayyaf Group, an Islamist terrorist organisation. This low-grade civil conflict continues despite a recent history of fairly functional multiparty democracy. Compounding ethnic and religious tensions are poor service provision in peripheral areas and a very gradual pace of land reform.

In recent work (Berman et al. 2016), we study the ‘Peace and Development Teams’ (PDT) programme, which was implemented by the Armed Forces of the Philippines between 2002 and 2010. PDT units worked in over 5,000 villages across the country, which account for 12% of the country’s population of nearly 100 million. The units were tasked with confronting rebel forces, securing the surrounding area, identifying local needs, and connecting villages with existing government services.

We focus on the one development outcome for which we have appropriate and reliable data in less governed areas of the Philippines – malnutrition of children aged 0-4. These data come from a large-scale government data collection initiative called Operation Timbang (for institutional reasons, we believe that data collection was conducted successfully even in remote, poorly governed spaces, and we find no evidence that the programme affected the quality of measurement). To estimate the effects of PDT on malnutrition we exploit the staggered roll-out of PDT across 712 of the Philippines’ 1,648 municipalities, drawing upon non-recipient municipalities to understand how malnutrition would have progressed in the absence of the program.

Results: Effects on malnutrition, violence, spending, and neighbours

Figure 1 presents our main results. It shows the estimated level of malnutrition during the five years before and seven years after implementation, all relative to the year before PDT (which is shown as a red diamond fixed at zero). The vertical lines represent the confidence interval, which accounts for variability and uncertainty in the estimates of how malnutrition progressed around the intervention. The dots and bars to the left of the figure (before the red diamond) show little evidence that malnutrition was already falling when PDT was implemented. That’s reassuring.  To the right of the diamond, however, is evidence that PDT set off a sustained decline in child malnutrition that levels out around three years later, and seems to remain for at least seven years. The magnitudes are quite large. The estimates imply that treating the entire municipality with PDT would lower malnutrition by as much as 30%.

This 30% long-term reduction is almost as large as the best nutritional interventions we know of, and bear in mind that those interventions would probably not be feasible in an ungoverned (or under-governed) space, for lack of security.

Why did malnutrition decline? Possible channels of influence include improving access to nearby fields and markets, improving health workers’ ability to work, or simply improving the safety of people and property by reducing rebel presence and accompanying extortion. While we cannot narrow it down completely, it is possible to rule out some competing explanations.

  • It was not declining violence.

In fact, measured violence rose in the months before, during, and after the programme. This rising violence is driven by government-initiated incidents as they contest rebels for control of the area. Thus, it is unlikely that the gains reflect ‘peace dividends’ among local children. At the same time, we find no evidence that PDT systematically increases deaths among children.

  • It was not government spending or federal involvement.

For most of the period, we have data on municipal government spending. We see no effects on either total spending or spending on health. While it’s difficult to observe all relevant types of federal involvement, if PDT were simply coincident with other investments in expanding state influence, we might expect an increase in federal transfers to recipient municipalities. We find no evidence of this. If anything, PDT was accompanied by declining federal-municipal transfers.

  • Insurgents are bad for child health.

Perhaps the best evidence for the mechanism by which PDT improved child health is from its perverse effects – rising malnutrition in neighbouring municipalities. A long literature has documented ‘displacement’ effects in conflict, in which rebel groups might simply be pushed to neighbouring municipalities (see Beardsley et al. 2015 for recent evidence on displacement in a large number of conflicts). A perennial concern in policy discussions about counterinsurgency is that an under-resourced programme might simply relocate, rather than reduce, rebels. Consistent with this, we find that PDT significantly increases malnutrition in neighbouring municipalities, an effect that is entirely concentrated among neighbours who never receive PDT themselves. This evidence is consistent with the interpretation that it is the presence of Maoist insurgents that is deleterious to child health, and that shifting the rebels simply shifts the problem. In fact, the spillover effects are so large that the aggregate program effects are near zero, since improvements among recipient municipalities are mostly (though not entirely) offset by worsening malnutrition among non-recipients.


Many of the world’s poorest people are concentrated in remote spaces, where governments are absent. Daesh-controlled space in Syria and Iraq is but one extreme example. A possible solution is expanding governance, but like any coercive act, that may increase risk to local populations. We provide evidence that a well-trained, targeted, and adequately resourced counter-insurgency programme can significantly improve welfare. Though we have direct evidence only on improvements in malnutrition, it’s reasonable to believe that improvements in security improved welfare in other ways as well. Moreover, that welfare improvement is evident even when the governance being expanded is relatively weak and the rebels are fairly innocuous (compared to Daesh or the Taliban).

Once governance is in place, several recent studies indicate that the provision of security enables effective development projects that might otherwise be destroyed, or looted, or even increase violence (Crost et al. 2014, Nunn and Qian 2014, Berman et al. 2013). While this does not suggest that military-centric strategies are a panacea for unstable countries, it does suggest that appropriate use of state forces should be part of the policy discussion. Conflict does enormous damage, but excluding populations from governance may be even worse.


Beardsley, K, K S Gleditsch and N Lo (2015) "Roving bandits? The geographical evolution of African armed conflicts", International Studies Quarterly, 59(3): 503-516.

Berman, E (2009) Radical, religious and violent: The new economics of terrorism, MIT Press.

Berman, E, M Downey and J H Felter (2016) “Expanding governance as development: Evidence on child nutrition in the Philippines”, NBER, Working paper w21849.

Berman, E, J H Felter, J N Shapiro and E Troland (2013) “Modest, secure, and informed: Successful development in conflict zones”, American Economic Review: Papers and Proceedings, 103(3): 512-517.

Crost, B, J H Felter and P Johnston (2014) “Aid under fire: Development projects and civil conflict”, The American Economic Review, 104(6): 1833-1856.

Heger, L L and D F Jung (2015) "Negotiating with rebels: The effect of rebel service provision on conflict negotiations", Journal of Conflict Resolution, 0022002715603451.

Keister, J (2011) States within states: How rebels rule, PhD thesis, UC San Diego.

Nunn, N and N Qian (2014) "US food aid and civil conflict", The American Economic Review, 104(6): 1630-1666.

315 Reads