Advanced analysis, anticipatory action and large-scale response

1Food crisis analysis

A comprehensive analysis cross-references several sources: household assessments, price and market monitoring, agricultural and pastoral data (rainfall, harvests, pasture conditions), nutrition, mortality, security and access. Livelihood profiles (HEA approach) help predict which groups will be affected and when.

It is shared within the IPC or Cadre Harmonisé technical working group and leads to projections. An NGO contributes its data and field knowledge, clearly stating their limitations.

2Anticipatory action

Acting before the peak of a foreseeable crisis costs less and protects better: transfers before the lean season, livestock destocking before animals lose condition, short-cycle seeds. Anticipatory action relies on triggers agreed in advance (rainfall, price or IPC projection thresholds), ready plans and pre-arranged financing.

Mechanisms exist: United Nations emergency funds, donors' anticipatory financing, climate insurance. The hardest part is often the decision: a plan with clear triggers avoids delays.

3Large-scale response and national systems

At scale, efficiency comes from common systems: beneficiary registration and deduplication, digital payments, shared price monitoring, harmonised transfer values. Shock-responsive social protection uses an existing national programme to assist more households (horizontal expansion) or increase transfer values (vertical expansion).

Data protection is essential: data-sharing agreements, minimisation, list security, informed consent of beneficiaries.

4Coordination, accountability and exit

The cluster sets geographic priorities, baskets, transfer values and common indicators; it tracks gaps and leads advocacy. Accountability is measured: rate of complaints resolved, satisfaction, women's participation in committees.

Every response plans its exit strategy: handover to safety nets, recovery of livelihoods, gradual phase-down. A poorly prepared exit pushes households back into crisis strategies.

5Integrated crisis analysis and scenario development

Advanced analysis of a food crisis is not limited to classifying the current situation: it seeks to anticipate what will happen, to whom and when, so that the response arrives before the deterioration. FEWS NET's scenario development methodology, reflected in the projection logic of the IPC and the Cadre Harmonisé, provides the structure.

The steps in developing a scenario

1. Define the area (livelihood zone or administrative unit), the projection period (generally six to eight months, divided into sub-periods) and the groups concerned.
2. Describe the current situation: household outcomes (consumption, livelihoods), nutrition, mortality, and the factors that explain them.
3. Set out the key assumptions about the driving factors: rainfall in the next season (seasonal forecasts), price trends (seasonal patterns, exchange rate, trade policy), conflict and access, planned and funded humanitarian assistance, population movements. Each assumption is dated and sourced.
4. Project the effect of these assumptions on each group's food and income sources, using HEA where a profile exists, or structured qualitative reasoning.
5. Project the outcomes: consumption, coping strategies, nutrition, likely phase for each sub-period.
6. Identify the events that would change the scenario (significantly better rains, conflict escalation, border closure) and their likely effect.
7. Monitor the assumptions and update the scenario as soon as an assumption no longer holds.

Early warning sources

Intersectoral analysis

In the Humanitarian Programme Cycle, the Joint Intersectoral Analysis Framework (JIAF) combines sectoral analyses to estimate the severity of needs and the number of people in need. The Food Security sector contributes through IPC or Cadre Harmonisé figures. The analyst must be able to explain the differences between IPC figures (populations in Phase 3 or above), the people in need in the HNO and the HRP targets: these are three different figures, built for different purposes.

A projection is only useful if it is linked to a decision. For each scenario, the analyst specifies what should be done, by whom and by what date, and what will happen if nothing is done. This is the bridge between analysis and anticipatory action.

6Designing an anticipatory action system

Anticipatory action means acting between the moment a shock becomes likely and the moment its humanitarian impacts peak. It has been extensively developed since the mid-2010s by the Red Cross and Red Crescent Movement (forecast-based financing, followed by the anticipatory pillar of the IFRC emergency fund), FAO, WFP, the Start Network and OCHA, with anticipatory action frameworks piloted in several countries and funded in particular by the Central Emergency Response Fund (CERF). Available evaluations, including several FAO studies, conclude that households assisted before the peak of a drought or flood suffer fewer asset losses and resort less to crisis strategies, often at a lower cost than a late response.

The four components of a system

1. A trigger model: which indicator, which threshold, which source, on what date, and who confirms that the threshold has been crossed.
2. Pre-agreed actions: what will be done, for whom, where, within what timeframe and with which partners.
3. Pre-arranged financing: reserved funds or a guaranteed mechanism (pooled fund, donor envelope, insurance), released automatically when the trigger is reached.
4. Operational readiness: pre-registered lists, framework agreements, payment service providers, prepositioned stocks, trained partners, early warning messages ready to go.

Choosing triggers

A good trigger is objective (whether it has been crossed is not open to debate), published by a reliable and independent source, available in time and demonstrably linked to the impacts to be avoided. It is analysed against historical data: how many times would it have been crossed over the past twenty years, and did those years correspond to real crises? A certain number of "false alarms" is accepted: acting needlessly one year in five can be cost-effective if major losses are avoided in the other years.

Choosing actions

Anticipatory actions must be achievable within the window between the trigger and the impact, and must protect what the shock will destroy: cash transfers before the lean season to prevent the sale of animals and indebtedness; commercial destocking while animals still have value; feed and veterinary care for the core breeding stock; short-cycle or drought-tolerant seeds; hermetic storage bags and moving stocks to safety before a flood; early warning messages with practical advice.

Financing instruments

• Central Emergency Response Fund (CERF): anticipatory allocations linked to anticipatory action frameworks approved in advance by OCHA and the agencies.
• IFRC Disaster Response Emergency Fund (DREF), anticipatory pillar: financing of Early Action Protocols (EAPs) prepared by National Societies.
• Start Network: the Start Fund (anticipation alerts) and risk financing mechanisms (Start Ready).
• Sovereign climate risk insurance, such as the African Risk Capacity (ARC), with implementation plans prepared in advance, and the mechanism that allows humanitarian organisations to take out parallel cover (ARC Replica).
• Donor envelopes: crisis modifiers and flexible funding lines built into contracts (ECHO, BHA, FCDO).

7Worked example: the anticipatory action plan for the northern pastoral zone

Fictional context: the northern pastoral zone of Haut-Kéla district is suffering increasingly frequent droughts. A consortium of three NGOs, working with FAO, the national livestock service and the national early warning system, is preparing an anticipatory action plan for 6,000 poor and very poor pastoral households. The rainy season runs from June to September; the critical dry season from February to May.

1. The trigger model

Retrospective analysis over twenty years shows that the combination of the vegetation index and terms of trade would have triggered activation six times; five of those years were indeed followed by a documented pastoral crisis, and only one crisis would have gone undetected. The consortium considers this level acceptable and documents it.

2. Actions, timing and cost

3. The economic argument

The consortium estimates the cost of a comparable late response: six months of food assistance (6,000 × 6 × 70 = USD 2,520,000) and partial restocking of herds (6,000 households × 5 goats × USD 60 = USD 1,800,000), i.e. USD 4,320,000, not counting income losses and effects on nutrition. The ratio is roughly 2 to 1. The consortium presents this figure as an internal estimate based on explicit assumptions, not as proof.

4. Financing and governance

• Readiness funds and 60% of the action budget reserved in a donor's crisis modifier, which can be activated on simple notification that the threshold has been crossed.
• The remainder requested under the national anticipatory action framework funded by a CERF allocation, with WFP and FAO.
• A technical committee (early warning system, livestock service, consortium, FAO) confirms within 72 hours that the thresholds have been crossed; the decision is not subject to renegotiation.

5. Monitoring and learning

The plan provides for a comparison between assisted households and non-assisted households with a comparable profile (neighbouring areas outside the programme) on: number of breeding animals retained, distress sales of animals, debts, and FCS and rCSI in March. The results are shared with the national anticipatory action working group and used to revise the thresholds and actions.

The hardest part is not technical: it is securing the prior commitment that the money will be released when the threshold is crossed, without renegotiation. An anticipatory action plan without pre-arranged financing is just one more contingency plan.

8Large-scale response and shock-responsive social protection

When several hundred thousand people need assistance, project-by-project approaches reach their limits: duplication, different transfer values, multiplied administrative costs, households registered five times by five organisations. ECHO's guidance on large-scale cash transfers, the UN Common Cash Statement (UNHCR, WFP, OCHA, UNICEF) and the Grand Bargain workstreams push towards common systems and towards the use of national social protection systems where this is possible and consistent with humanitarian principles.

Options for using a national system

Worked figures

In the fictional northern region, the national safety net programme pays 15,000 CU per month to 15,000 households. The social registry contains 42,000 registered households, of which 27,000 are not enrolled. The response plan combines: a vertical expansion of 39,000 CU per month for four months for the 15,000 enrolled households (bringing the total to 54,000 CU, the CWG amount), i.e. 2.34 billion CU; a horizontal expansion to 10,000 households from the registry who meet the criteria of heightened vulnerability, at 54,000 CU for four months, i.e. 2.16 billion CU; and an aligned humanitarian response for displaced people, who are not in the registry. Payments go through the national programme's provider, and NGOs handle verification, accountability and monitoring in their areas.

Common systems

• Registration and deduplication: a unique identifier, or at least a comparison of lists between organisations according to agreed rules (matching on name, date of birth, phone number, location, or biometrics where proportionate and lawful).
• Payments: one or two jointly selected providers, negotiated fees, interoperability between platforms.
• Segregation of functions: the organisation that targets is not the one that pays or the one that verifies, to reduce the risk of fraud.
• Common monitoring: harmonised PDM, common indicators, shared dashboard.

Data responsibility

The larger the system, the higher the risk: a beneficiary list can expose people to reprisals or commercial exploitation. The ICRC Handbook on Data Protection in Humanitarian Action and the IASC Operational Guidance on Data Responsibility in Humanitarian Action set out the principles: a legitimate basis and informed consent, data minimisation, specified purpose, security, limited retention period, written data-sharing agreements, and an impact assessment before any new data collection or sharing. Sharing lists with authorities that are party to a conflict, or with commercial providers beyond what is strictly necessary, requires an in-depth risk analysis and may be refused.

9Prioritising, coordinating and exiting: deciding when resources fall short

In most protracted crises, funding covers only a fraction of needs. The question is no longer just "who needs assistance?" but "who should be assisted first, with what level of assistance, and how can this be justified?". The cluster plays a central role here: prioritisation decided by each organisation on its own produces inconsistencies and tensions.

A four-level prioritisation method

1. Geographical prioritisation: areas classified in Phase 4 or above and areas with the largest populations in Phase 4; areas with a "!" where stopping aid would cause deterioration.
2. Household prioritisation: criteria common to the cluster, from most to least vulnerable (for example CARI 4, then CARI 3 with emergency strategies, then CARI 3), with automatic inclusion categories.
3. Level of assistance: a full ration for the most severe cases, a partial ration for the others, according to measured deficits (HEA or expenditure).
4. Duration and timing: concentrate resources on the lean season rather than spreading amounts that are too small over the whole year.

The decision to cut rations or transfer values must be taken collectively, communicated to communities in advance, accompanied by closer monitoring (consumption, coping strategies, nutrition, protection) and reversed as soon as funding allows.

Gap analysis and advocacy

Using the 5W, the cluster maintains a gap analysis by area: population in need, population targeted by the plan, population actually assisted, level of assistance received. Effective advocacy draws on these figures, on the official classification, on the cost of inaction (projected deterioration, cost of a late response) and on concrete, costed proposals, rather than on a general appeal.

Exit strategy and transition

Measuring accountability across the response

The cluster can track a few common indicators: proportion of households who know their entitlements and complaint channels, average complaint handling time, proportion of complaints that received a response, participation of women in committees, results of the Cluster Coordination Performance Monitoring (CCPM), and perceived fairness of assistance as measured in PDM.

10Variations by context, common mistakes and donor expectations

Variations by context

Common mistakes

• Designing a trigger without checking its historical performance, or basing it on data published too late to act.
• Planning actions that cannot be carried out within the window between the trigger and the impact.
• Leaving the activation decision to be negotiated at the time of the trigger.
• Using a social registry without checking that it is up to date (deceased or relocated households, newly poor households missing).
• Sharing lists without an agreement or a risk analysis.
• Cutting rations without informing communities or monitoring the effects.
• Exiting a programme on the date funding ends rather than according to criteria.

What donors, auditors and evaluators look for

• ECHO: integration of preparedness and anticipatory action (crisis modifiers, disaster preparedness), segregation of functions and cost-efficiency in large-scale cash programmes, data protection, links with social protection within a nexus approach.
• BHA: justification of transfer value and modality, market monitoring plans, use of the crisis modifier, coordination with national mechanisms and the clusters.
• FCDO and development donors: contribution to strengthening national social protection systems, evidence on the cost-effectiveness of anticipatory action.
• Evaluators: OECD DAC criteria (relevance, coherence, effectiveness, efficiency, impact, sustainability), with particular attention to timeliness (did the aid arrive before the peak?) and connectedness (did the programme prepare for what comes next?). ALNAP's guide on evaluating humanitarian action provides the reference framework.

11Interview questions and key references

Interview questions

1. "Design an anticipatory action system for a drought-prone area in five minutes."

A good answer cites the four components (trigger, pre-agreed actions, pre-arranged financing, readiness), proposes combined triggers (seasonal forecast, vegetation, terms of trade, IPC projection), actions suited to the pastoral or agricultural calendar, a financing instrument, and a monitoring system to test the thresholds.

2. "What is the difference between vertical and horizontal expansion, and when would you choose one or the other?"

The answer defines both, explains that vertical expansion is fast but only reaches enrolled households, and that horizontal expansion requires an up-to-date registry and enrolment capacity, then shows that the two are often combined with a humanitarian response for people missing from the registries.

3. "Funding covers 40% of needs. How should the cluster prioritise?"

The answer describes prioritisation first by geography, then by household, then by level of assistance and timing, decided collectively, communicated to communities, monitored and revisable, accompanied by costed advocacy.

4. "A ministry asks for access to your beneficiary lists to include them in the social registry. What do you do?"

The answer analyses the legal basis, people's consent, the purpose and the risks (particularly in a conflict setting), proposes a data-sharing agreement with minimisation and security, consults the data protection and protection leads, and may propose alternatives (aggregated sharing, voluntary enrolment).

5. "How do you demonstrate that an anticipatory action had an effect?"

The answer proposes a comparison with households of a comparable profile that were not assisted or were assisted later, asset indicators (animals retained, distress sales, debts) and consumption indicators, measurement at the time the shock peaks, and an analysis of avoided costs, while acknowledging the methodological limitations.

6. "What exit strategy do you plan for an emergency cash transfer programme?"

The answer mentions exit criteria set at the design stage, the seasonal calendar, a gradual approach, information for communities, handover to social protection and livelihoods programmes, and post-exit monitoring.

Key references

• Global Report on Food Crises: Global Network Against Food Crises.
• Hunger Hotspots reports: FAO and WFP.
• Scenario development guidance and food security outlooks: FEWS NET.
• Documents on anticipatory action frameworks and CERF anticipatory allocations: OCHA.
• Manual and resources on forecast-based financing and Early Action Protocols: International Federation of Red Cross and Red Crescent Societies and the Red Cross Red Crescent Climate Centre.
• Resources on anticipatory action: FAO; Start Network; Anticipation Hub.
• Work on shock-responsive social protection: Oxford Policy Management, World Bank, Socialprotection.org.
• ECHO guidance note on large-scale cash transfers; UN Common Cash Statement.
• Handbook on Data Protection in Humanitarian Action: ICRC; Operational Guidance on Data Responsibility in Humanitarian Action: IASC.
• Reference Module for Cluster Coordination at Country Level: IASC; Humanitarian Programme Cycle and Joint Intersectoral Analysis Framework documents.
• ALNAP guide on evaluating humanitarian action.