Identifying vulnerable households — and what to do after you have

Nearly every RAP contains a definition of vulnerability and a commitment to provide additional assistance to vulnerable households. Rather fewer contain any mechanism by which a household identified as vulnerable at census receives anything different afterwards. The flag is set and never read.

Vulnerability is about capacity to recover, not category

The standard formulations list characteristics — elderly, disabled, female-headed, child-headed, landless, chronically ill, indigenous, very poor.[1] Lists are useful for enumeration and misleading as an analysis, because the operative question is not which box a household falls into but whether it can absorb a shock and reconstruct a livelihood afterwards.

Two households can share every listed characteristic and differ entirely: one has an employed adult child nearby, the other does not. The characteristics are proxies for capacity, and treating them as the thing itself produces a list that includes households that will be fine and misses households that will not.

What to collect at census

Practical additions to the standard instrument, none of them expensive:

• Dependency structure — how many people rely on the household's income, and how many contribute to it.
• Severity of loss as a proportion of the household's total productive base, not in absolute area. A household losing all of a small holding is more exposed than one losing a larger area from a bigger holding.
• Income concentration — a single source, or several.
• Documentation status — whether adults hold identification, which determines whether they can be paid at all.
• Support network — whether relatives nearby could absorb a temporary loss of income.

Severity of loss deserves particular emphasis. It is the strongest single predictor of impoverishment, it is derivable from data the census already collects, and it is almost never used as a vulnerability criterion.[2]

A vulnerability flag that does not change what happens to the household is a documentation exercise. The question to ask of any criterion is: what does this trigger?

Making the flag operational

For the flag to matter it has to be attached to specific, funded actions with named owners. The ones that reliably help:

1. Assisted process. Someone accompanies the household through enumeration verification, entitlement disclosure, documentation and payment, rather than expecting it to navigate an administrative process alone.
2. Priority in the payment queue, since delay falls hardest on households with no reserves.
3. Documentation support — obtaining identification, letters of administration, or succession documents, started at census rather than at payment.
4. Physical assistance with the move, which for an elderly or disabled household is not a convenience but the difference between moving and not.
5. Transitional support covering the gap between displacement and the replacement asset becoming productive.
6. Individual monitoring through implementation and for a defined period after, at household level rather than in the aggregate.

Payment risk

A large payment to a household with no experience of holding one is itself a risk, and it falls hardest on exactly the group the flag identifies: elderly recipients, widows, and households where the registered claimant is not the person who managed the productive asset.[3]

Available mitigations include staged payment tied to replacement asset purchase, accompanied banking, and — where the household consents — payment to more than one named recipient. All of these carry a paternalism cost that has to be weighed honestly, and all of them require the household's informed agreement rather than an administrative decision made on its behalf.

Tracking through to closure

Vulnerable households are the ones a completion audit will sample, and they are the ones most likely to have moved, changed composition or lost contact by the time it arrives. So the register needs to carry them individually: what was flagged, what assistance was committed, what was delivered, when, and what the household's status was at the last contact.[4]

Aggregate reporting — a count of vulnerable households assisted — cannot answer the question an auditor asks, which is about a named household on a named date. That record either exists contemporaneously or it does not exist at all.