Assessing vulnerability at the household level has become a central question since the start Syrian refugee crisis. The giant toll placed on countries and humanitarian agencies from the sheer number of the displaced (now over 4.5 million estimated by UNHCR) has naturally led these actors to nuance their assistance, targeting only those most in need. There is a consensus in humanitarian agencies such as the World Food Programme and the UN Refugee Agency that blanket support for refugees past their first months of displacement is very costly and leaves a rather large inclusion error.
Based on this, attempts have been made to target only the most vulnerable and food-insecure refugees for assistance, which ranges from the distribution of tools and shelter to the monthly provision of food.
The issue is that different offices and operations have come to very different conclusions about what vulnerability is, how to measure it, and how to use that measurement to target only a portion of the refugees.
Of course different sectors are trying to achieve different objectives- ensuring that refugees are food-secure, or that they have enough income, or safe access to water and sanitation, and so on. But even within the food security sector, the plethora of methods used to decide who is vulnerable and who is not reflect a lack of consensus on targeting methodology.
To be sure, humanitarian agencies are good at targeting: WFP does a lot of geographic targeting, identifying in a population spread out across an area, where those most likely to be food-insecure are living. UNHCR also does targeting based on a short list of socio-demographic criteria- identifying all single women heads of household, or all households with a disabled/chronically ill/elderly person, and so forth.
The complicated nature of targeting in the Syrian case (not just Syria of course- something similar is being done in Sudan) is due to the requirement of having a by-household assessment of food security/vulnerability. This cannot be done by only looking at a sample of households. Geographic spread loses some importance, and single-variable filters (presence or absence of female as head of household) do not capture the multi-dimensional nature of vulnerability.
Therefore the ‘by-household’ part precludes in large part anything but a census to identify households. The ‘food security’ part, while pretty much straightforward, poses problems when it comes to asking ‘how food secure?’
The targeting has been mixed across Jordan, Lebanon and Egypt. It has included a proxy-means test in some locations, a combination of a means-test on target indicators (consumption, income, etc.) and a ‘score card’ approach in others, and even a sort of cross-tabulation of food consumption and coping strategy indices, and poverty indicators.
The mixed nature of this approach is probably more due to differences in thinking across offices than to anything else. The problem is that there is almost no way to compare how well each approach is vis-à-vis the other. It is possible to calculate the extent of inclusion and exclusion errors (not always easy!), but each of these approaches actually measures a different thing- there are several ways to measure food security, and differences in the questions asked and the weights given to different factors means there is an intrinsic barrier to comparison.
Sudan is another issue, since the population is not one of freshly displaced refugees, but one of IDPs displaced 10+ years ago. The IDP identity is highly politicized in Sudan, and the registration of IDPs that took place 10 years ago, along with the subsequent caseload verification, are widely acknowledged in the humanitarian community in Sudan to have been fraudulent to a certain extent. Furthermore there are serious issues of access and security in certain locations in Darfur which may not exist in refugee host countries. The Sudanese case, where targeting methodologies are different still, is perhaps more complex for these reasons.
A couple of parting thoughts on identifying and targeting households by vulnerability, from my experience and conversations with colleagues:
- Context is important– where are people living? How is their access to services, community assets, connection to the host country? How restricted is their movement and what is their rate of participation in the local economy? All these factors look beyond the single household, but can greatly inform how different locations are treated differently.
- Vulnerability, and measuring vulnerability, is subjective– every methodology has its faults, and, operationally speaking, there will always be the fear of exclusion errors. Resources permitting, it is better to err on the side of inclusion rather than exclusion.
- Beneficiary feedback and cross-checking can help– the most glaring exclusion errors can be caught with a good system to process complaints. People who were pushed out of assistance through the targeting should be able to register a complaint and have their case re-examined, time and resources permitting.