Cash transfers at scale: risks, data and social protection

1Monitoring markets and adjusting values

A cash programme lives with markets. Price monitoring (at least monthly, often joint within the cash working group) makes it possible to recalculate the MEB and adjust transfer values according to a mechanism agreed in advance: trigger threshold (for instance +15% on the basket), review frequency, harmonisation between organisations to avoid gaps between neighbours.

Post-distribution monitoring (PDM) measures how transfers are used, access to markets, risks encountered and effects (food consumption, coping strategies, debt).

2Risks: fraud, security, data

Risks are managed through a cash-specific risk register: diversion by intermediaries, ghost recipients, collusion with the provider, insecurity at pay points, exposure of personal data.

• Controls: payment reconciliation, random verification, an independent complaints line, contract clauses with the provider.
• Data: data protection impact assessment, minimisation, written data-sharing agreements, encryption and limited retention. Displacement status, ethnicity or religion are sensitive data.

3Towards social protection

In protracted or recurrent crises, humanitarian cash can build on national social protection systems or prepare a transition: using social registries, aligning values and criteria, supporting shock-responsive mechanisms (vertical or horizontal expansion of existing programmes).

This link must preserve humanitarian principles: where the State is a party to the conflict or certain populations are excluded, humanitarian aid keeps its independence.

4Adjusting transfer values: indexation, inflation and exchange rate risks

A transfer value set in the reference month loses its relevance as soon as prices or the currency change significantly. In protracted crises, particularly in high-inflation economies, adjusting values becomes a regular management activity, which must be planned from the design stage and coordinated within the CWG.

Components of an adjustment mechanism

Difficult decisions

• Falling prices: transfer values are rarely reduced, but a sustained fall should lead to a revision, otherwise the gaps between programmes widen. The rule must be symmetrical from the outset.
• Fixed budget: if the value increases without additional funding, the programme must reduce the number of households or the duration. This decision must be explicit, discussed with the donor and announced to communities; keeping a value artificially low amounts to silently reducing assistance.
• Gaps between organisations: an organisation that revises its values on its own creates tensions. Harmonisation within the CWG takes precedence, except in documented emergencies.

Payment currency and exchange rate risks

In economies where the local currency is depreciating rapidly, where several exchange rates coexist or where local currency liquidity is scarce, the choice of currency becomes a strategic decision.

Where there is a large gap between the official rate and the parallel market rate, converting funds at the official rate can sharply reduce the real value received. The decision involves finance, senior management, the donor and often the Humanitarian Coordinator, as it touches on national regulations and relations with the central bank. Exchange gains and losses must be monitored and reported according to the donor's rules.

The right indicator is not the nominal amount paid, but the household's real purchasing power: what share of the MEB can the transfer buy at the time the household spends the money?

5Fraud, diversion and sanctions: a control system commensurate with scale

The larger a programme, the more fraud attempts it attracts, and the more costly a single weakness can be. Cash transfers are not inherently more exposed than in-kind assistance (food is also diverted, but in a less traceable way), but their risks are different and require controls adapted to each stage.

Mapping risks by stage

Red flags in the data

• Several households registered with the same phone number, the same identity document or the same GPS coordinates.
• Abnormally short interview durations, or forms all completed at the same time by the same enumerator.
• Vulnerability scores clustered just above the eligibility threshold, a possible sign of manipulated answers.
• Withdrawals made within a few minutes for many accounts at the same agent.
• A total absence of complaints in one area while other areas receive many.
• An abnormally high percentage of households served on the very first day in an area where access is difficult.

What to do in case of suspicion

1. Protect: secure the evidence (files, logs), protect the people who reported the facts, do not confront the suspected person.
2. Report internally in accordance with the organisation's policy (often to the compliance officer or a dedicated hotline), within the required timeframe.
3. Inform the donor: most funding agreements require prompt notification of any suspected fraud or diversion, often within a few working days; check the applicable clause.
4. Investigate through a person or team independent of the programme and trained in investigations.
5. Correct: disciplinary measures, recovery, adjustment of controls, reimbursement of affected households.
6. Learn lessons: update the risk register and procedures.

Sanctions and countering the financing of terrorism

Organisations are subject to the applicable sanctions regimes (United Nations, European Union, national legislation and, depending on donors, other regimes) and to the clauses of their funding agreements. In practice, this means screening providers, partners, suppliers and staff against the applicable lists. Screening of the beneficiary households themselves is generally not required by humanitarian donors, and many organisations refuse it as contrary to the principle of impartiality: what each agreement requires must be checked precisely. United Nations Security Council resolution 2664 (2022) introduced a cross-cutting humanitarian exemption to UN sanctions regimes, without removing the obligations set out in other regimes or in funding agreements.

The objective is not zero risk, which would lead to no longer serving hard-to-reach areas, but a known risk, reduced to an acceptable level, explicitly shared with the donor and partners, and monitored.

6Data at scale: interoperability, requests from authorities and incidents

At scale, cash transfer programmes generate databases of tens or hundreds of thousands of households. These databases are of interest to other organisations (to avoid duplication), to authorities (for planning or control), to providers (for their own commercial activities) and sometimes to malicious actors. The issue is no longer just protecting a database, but governing data flows.

Interoperability and de-duplication between organisations

• Principle: share the minimum necessary to achieve the purpose (avoiding duplicates, referring cases), not entire databases.
• Techniques: transformed (hashed) identifiers that make it possible to identify a duplicate without revealing identity; sharing limited to an "already assisted by another organisation" flag; common platforms managed by a designated entity, with written access rules.
• Governance: data sharing agreement signed by all parties, data governance committee within the CWG, annual review.

Responding to a request from the authorities

1. Acknowledge receipt and ask for the precise purpose and the legal basis of the request.
2. Analyse the risks to people: could the data expose them to arrest, discrimination, forced recruitment or eviction?
3. Check the commitments made to households when obtaining consent, and the organisation's policy.
4. Propose alternatives: data aggregated by area, anonymised statistics, individual referral only with the person's agreement.
5. Escalate to the country director and, if necessary, to headquarters; coordinate with other organisations, through the CWG and humanitarian coordination, for a common position.
6. Document the decision and the response given.

Biometrics: a choice that must be justified

Biometrics (fingerprints, iris, facial recognition) allow effective de-duplication at very large scale and reduce certain types of fraud. But biometric data cannot be changed if compromised, can be reused for other purposes, and its collection can be perceived as a condition imposed for receiving assistance. The ICRC has adopted a specific policy that strictly regulates its use. The decision requires an impact assessment, a demonstration that less intrusive means are not sufficient, safeguards on storage (preferably on a card held by the person rather than in a central database) and an alternative for people who refuse or whose fingerprints cannot be read.

Managing a data incident

A humanitarian database can outlive the programme, the organisation and the political context in which it was created. Collect and retain data with the worst possible use a hostile actor could make of it in mind.

7Linking humanitarian assistance with social protection: options, conditions and method

In recurrent or protracted crises, millions of households receive humanitarian cash assistance every year while the state runs, or is beginning to build, social transfer programmes. Linking the two makes it possible to reduce costs, avoid parallel systems, prepare an exit from humanitarian aid and respond more quickly to subsequent shocks. The CALP Network, the Grand Bargain sub-group on linking humanitarian cash and social protection, and the Social Protection Inter-Agency Cooperation Board (SPIAC-B) have produced guidance on this subject.

Typology of options

A widely used typology, drawn from Oxford Policy Management's work on shock-responsive social protection, distinguishes five ways of using or mirroring a national system:

Assessing the national system: questions to ask

• Coverage: what share of affected households is listed in the registry or benefits from a programme? Are displaced people, refugees or minority populations included?
• Registry quality: date of the last update, targeting method, known error rates.
• Payment mechanisms: reliability, timeliness, fees, coverage of affected areas.
• Values: social transfers are often well below the gap measured by the humanitarian MEB; convergence requires discussing this gap.
• Legal framework and data: legal basis for data sharing, data protection, redress.
• Principles: is the state a party to the conflict? Are some populations excluded for political reasons? Would using the system compromise the neutrality, impartiality or independence of humanitarian action?

Working method

1. Map national programmes, their components and the development partners that fund them (often the World Bank, UNICEF, WFP, ILO).
2. Establish a regular dialogue between the CWG and the ministry responsible for social protection, with a focal point on each side.
3. Choose one or more options by area and population, based on the analysis; it is common to use vertical expansion for households already enrolled and an aligned humanitarian programme for populations not covered.
4. Harmonise progressively: targeting criteria, payment calendar, values, complaints mechanisms, compatible registration forms.
5. Prepare the data: data sharing agreement, transfer of the strict minimum, household consent, exclusion of sensitive data.
6. Plan the transition: timeline, criteria for transfer, communication with households, monitoring of transferred households for at least two cycles.

Social protection is not an automatic exit strategy. Households transferred to a national programme often receive a lower and less frequent value. The transition must be explained, monitored and accompanied by advocacy for the adequacy of national values.

8Worked example: 12,000 households in Kalandor, 30% inflation, fraud and transition

Fictional context: in Varno province (in the fictional country of Kalandor), the NGO "Réponse Commune Varno", in a consortium with two national NGOs, assists 12,000 vulnerable displaced and host households through monthly multipurpose transfers, funded for twelve months. The value is set at 75% of the SMEB for a six-person household, in line with the CWG recommendation. The local currency is the Kalandor dinar (DK). The budget is in US dollars, and an adjustment reserve of 5% of the annual value of transfers has been planned.

Situation in month 4

The review rule provides for a review if the SMEB rises by more than 15% for two consecutive months: it has been triggered since month 3. The CWG technical group recommends restoring the value to 75% of the updated SMEB.

Recalculation and budgetary effect

• New value: 75% of 3,120 DK = 2,340 DK per month.
• In dollars, at the month 4 rate: 2,340 divided by 25 = USD 93.60, compared with USD 90 at the start. The depreciation of the dinar absorbs most of the price increase: the cost in dollars rises by only 4%.
• Monthly cost: 12,000 multiplied by 93.60 = USD 1,123,200, compared with USD 1,080,000 planned, i.e. an additional cost of USD 43,200 per month.
• For the remaining eight months (months 5 to 12): additional cost of USD 345,600, if the exchange rate and prices remain stable.
• Available reserve: 5% of 1,080,000 multiplied by 12, i.e. USD 648,000. The reserve covers the additional cost and leaves a margin of USD 302,400 for further changes.

The team presents the donor with a risk scenario: if the exchange rate returned to 20 DK to 1 USD with stable prices, the monthly cost would rise to USD 1,404,000, i.e. an additional cost of USD 324,000 per month, which would exhaust the reserve in two months. It obtains the donor's agreement for a monthly review of the value and for a rapid amendment clause in the agreement if the reserve falls below 20% of its initial amount.

The fraud case

The month 5 PDM reveals that, in the districts of Hora and Bélen, 1,850 households report having had to hand over about 10% of their withdrawal (234 DK out of 2,340) to "facilitators" present near the payment agents. Analysis of the transactions shows bulk withdrawals, within a few minutes, at three agents. The team applies the procedure: internal report and report to the donor within the contractual deadline, investigation entrusted to the compliance unit, written notification to the provider, suspension of the three agents by the provider, deployment of supervised mobile agents, and a messaging campaign reminding households that the service is free and that their PIN is confidential. The affected households are reimbursed by the provider in the next cycle, i.e. 1,850 multiplied by 234 DK, about 432,900 DK. The risk register is updated and an automatic analysis of bulk withdrawals is added to each reconciliation.

The data request

The Ministry of the Interior requests the full database, including displacement status and district of origin. The team, together with the CWG, offers data aggregated by district and by month, refuses to transmit displacement status and district of origin at individual level, and offers named referrals only for households that request them in order to access a service. The common position is carried by the Humanitarian Coordinator.

The transition to the national programme

The government launches a national social safety net programme in the province, with a social registry already covering 40% of districts and a value of 1,500 DK every two months. The consortium prepares a three-phase plan:

Monitoring of transferred households over two cycles shows that 9% did not receive the first national payment, mainly because of identity errors. These cases are referred to the ministry through the agreed channel and the humanitarian top-up is maintained until the situation is regularised.

At each decision, the team asked itself the same question: what best protects the real value received by households and their safety, while remaining defensible before the donor, the CWG and the authorities?

9Contextual variations, common mistakes and expectations at scale

Adapting the system to the context

Common mistakes

• Not planning a reserve or an adjustment rule: when inflation strikes, the only remaining option becomes a silent reduction in purchasing power.
• Thinking in nominal amounts rather than purchasing power, and ignoring the effect of the exchange rate on a budget in foreign currency.
• Treating fraud as a taboo subject: the organisations that never detect anything are not the most honest, but those that control the least. Donors value prompt reporting accompanied by corrective action far more than silence.
• Transmitting entire databases to partners or authorities without a data sharing agreement or a risk analysis.
• Treating the transition to social protection as a simple transfer of lists, without monitoring transferred households or analysing the adequacy of values.
• Forgetting the humanitarian principles in the dialogue with the state: technical alignment must not compromise impartial access to assistance.

What donors expect for large programmes

• A common approach: in its guidance on large-scale cash transfers, DG ECHO encourages programmes designed to pool registration, the delivery mechanism, monitoring and accountability, with a segregation of functions between organisations to strengthen controls.
• Measured efficiency: ratio of value transferred to total cost, cost per household, comparison with similar programmes; several donors use cost-efficiency analysis tools based on actual expenditure data.
• Shared risk management: updated risk register, incident notification procedure, control plan.
• A long-term vision: exit strategy, linkages with national systems, contribution to preparedness for future shocks (framework agreements, forecast-based anticipatory financing).

At scale, the quality of a programme is judged less by the perfection of each procedure than by its ability to detect quickly what is not working and to correct it without losing the trust of households, communities and donors.

10Interview questions and key references

Interview questions

1. "Prices have risen by 30% in three months and your budget is fixed. What do you propose?"

A good answer refers to the planned review rule, recalculates the MEB or SMEB, analyses the effect of the exchange rate on the budget in foreign currency, uses the reserve, presents the donor with costed scenarios and the possible trade-offs (value, coverage, duration), and stresses harmonisation within the CWG and communication with households.

2. "How would you design an anti-fraud system for a programme covering 50,000 households?"

The candidate maps risks by stage, emphasises segregation of duties, cites preventive and detective controls (de-duplication, reconciliation, transaction analysis, verification calls, PDM, complaints mechanism), describes the procedure in case of suspicion and mentions the obligations to notify the donor. They raise the issue of proportionality: overly heavy controls exclude the most vulnerable households.

3. "Should beneficiaries be screened against sanctions lists?"

A strong answer distinguishes the screening of providers, partners, suppliers and staff, which is generally required, from the screening of beneficiaries, which is rarely required and, for many organisations, contrary to the principle of impartiality. It mentions the humanitarian exemption in Security Council resolution 2664 and the need to read each funding agreement carefully.

4. "Explain the difference between vertical expansion and horizontal expansion."

The candidate defines both options and their conditions (existing programme, up-to-date registry), gives an example and cites the other options in the typology (piggybacking, shadow alignment, refocusing). They mention the limitations related to humanitarian principles in conflict settings.

5. "A mobile network operator offers to manage payments free of charge in exchange for access to beneficiaries' data. What do you think?"

A good answer rejects the commercial reuse of household data, explains the risks (targeted marketing, indebtedness, exposure), recalls the purpose limitation principle and the need for a contract that prohibits such reuse, and proposes a normal competitive tender. It may mention that households can freely open accounts with the operator, on their own decision and after the humanitarian relationship has ended.

6. "How do you prepare the end of a cash transfer programme in a protracted crisis?"

The candidate discusses the analysis of transition options (social protection, livelihoods, "cash plus"), advance notice to households, a gradual phase-down if justified, post-programme monitoring, referral to other services and coordination with the CWG to avoid abrupt cut-offs for the most vulnerable.

Key references

• CALP Network: Programme Quality Toolbox, resources on linking cash assistance and social protection, The State of the World's Cash reports.
• Oxford Policy Management: research on shock-responsive social protection and the associated typology.
• SPIAC-B (Social Protection Inter-Agency Cooperation Board): guidance notes on social protection in crisis contexts.
• DG ECHO: thematic policy document on cash transfers and guidance note on large-scale cash transfers.
• IASC Operational Guidance on Data Responsibility in Humanitarian Action; ICRC Handbook on Data Protection in Humanitarian Action; ICRC Biometrics Policy.
• United Nations Security Council resolution 2664 (2022) on the humanitarian exemption to sanctions regimes.
• IASC Cash Coordination Model (2022).
• Grand Bargain commitments on cash transfers and linkages with social protection.
• Sphere Handbook (2018) and Core Humanitarian Standard (CHS).