The Case for Anticipatory Cash

In Ethiopia between 2022 and 2023, I was consulting for UNICEF cash transfer program and the Ethiopia Cash Working Group as drought was tightening across the Somali, SNNP, and Oromia regions. The forecasts were clear: rainfall deficits were widening, CHIRPS data showed below-average precipitation across multiple consecutive seasons, and the Standardised Precipitation Index was crossing thresholds we associated with mid-severity drought.

What we did not do, despite the forecasts, was release multi-purpose cash transfers in advance of the worst impacts. The funding mechanisms were post-onset. The humanitarian appeals were post-impact. By the time the CERF-funded MPC programme I monitored reached 185,000 beneficiaries with USD 3.2 million distributed, livelihoods had already been depleted, livestock had been sold at distress prices, and households had taken on debt to bridge the gap.

The economic literature on this is clear: anticipatory cash, deployed weeks before peak impact, costs a fraction of post-onset cash and produces materially better outcomes. The World Food Programme estimates that anticipatory action can save up to USD 5 for every USD 1 invested. The Anticipation Hub documents case after case where pre-positioned cash transfers prevented the cascading losses that post-onset response had to mop up.

So why isn't anticipatory cash the default mode of humanitarian response for slow-onset disasters? The answer is institutional, not technical. The technology to forecast and trigger has existed for years. What's missing is the data architecture to authorise.

Cash is uniquely suited to anticipatory action for three reasons that other modalities can't match.

Fungibility. A household receiving USD 50 in advance of a drought's peak impact can use it to buy fertiliser to extend the planting season, repair a borehole, send a child to relatives in a less-affected area, or stockpile staple foods. The same USD 50 in food rations can do exactly one of those things. The flexibility of cash is what makes it preventive — recipients deploy it against the specific risk their household faces.

Speed. Mobile money rails — M-Pesa, SafeBoda, regional bank transfers — can move funds in hours once the trigger fires. Procurement, shipping, and distribution of in-kind aid takes weeks. When the forecast tells you a flood will hit in 14 days, only cash can get to households in time to enable preventive action.

Dignity. Anticipatory cash treats recipients as economic actors making decisions about their own household risk. In-kind anticipatory aid, by contrast, is the humanitarian system telling people what they need before the disaster has even arrived. The accountability literature is unambiguous on which approach earns more trust.

Here is what changed in the last five years: the technical components of an anticipatory cash trigger are all in place.

Hazard forecasts. CHIRPS for rainfall, GloFAS for river discharge, NDVI and Vegetation Health Index for crop stress, sea-surface temperature for cyclone formation. These data streams are continuous, open, and global.

Impact baselines. DELTA Resilience — the next-generation national disaster loss tracking system co-developed by UNDRR, UNDP, and WMO — provides the historical impact records that turn a hazard forecast into an impact-based trigger. The forecast tells you the rainfall deficit will hit a specific threshold. The DELTA loss records tell you that the last three times the deficit hit that threshold in this district, it displaced X thousand people, destroyed Y hectares of crops, and pushed Z health facilities into overwhelm.

Vulnerability layers. Pre-positioned vulnerability indices, like the social vulnerability index I built for Newcastle City Council during my MSc, identify the populations most at risk before the event. In Afghanistan, my team produced national flood, drought, avalanche, landslide, and extreme-temperature risk mapping at 4 km resolution that identified 223 of 401 districts in extreme to abnormally dry conditions.

Disbursement infrastructure. Mobile money, agent banking, and cash-in-hand mechanisms now reach the majority of vulnerable populations in most operating contexts. The CERF-funded MPC programme in Ethiopia disbursed USD 3.2 million across three regions through four implementing partners — the rails exist.

The pieces are all there. The data tells us when to act, who is at risk, and how to move the money. What's missing is the institutional decision to trigger before the event rather than after.

The donor problem. Most humanitarian funding mechanisms require evidence of impact before disbursement. CERF Rapid Response works post-onset. Country-based Pooled Funds work post-onset. Bilateral appeals are launched in response to declared emergencies. The architecture is reactive by design, and asking it to be anticipatory requires re-engineering risk appetites that have been calcified for decades.

The trigger problem. A pre-agreed trigger — "cash releases when CHIRPS rainfall deficit exceeds 1.5 standard deviations AND IPC food security classification reaches Phase 3 AND historical DELTA loss records indicate displacement greater than 5,000 in this zone" — requires negotiation between donors, implementing partners, governments, and forecasters. The negotiation takes months. The drought doesn't wait.

The verification problem. Donor accountability frameworks were built around proving impact post-disbursement. Anticipatory cash is by definition disbursed before the impact materialises. Demonstrating value-for-money requires comparing what happened to a counterfactual where the cash wasn't released. That's methodologically harder than standard impact evaluation, and donor evaluation departments are still building the muscle.

The political problem. Releasing money for a disaster that hasn't yet happened looks, to a sceptical observer, like premature spending. If the forecast turns out wrong, the post-mortem is brutal. If the forecast was right and the cash prevented the worst impacts, there's no headline because nothing visible happened. The political incentives reward post-onset response over pre-onset prevention.

None of these are technical problems. All of them are solvable. But they require humanitarian leadership willing to take the institutional risk that the system structurally discourages.

Drawing on the Bangladesh shock-responsive cash work that informed USD 45.5 million in transfers reaching 2.5 million vulnerable people, on the Ethiopia CWG architecture, and on the DELTA Resilience framework I've been writing about — here is what a mature anticipatory cash system requires.

Pre-agreed triggers, pre-positioned funding, pre-vetted implementing partners. All three have to be in place before the event. Negotiating any of them in the 14-day window between forecast and impact guarantees the cash arrives late.

A multi-source forecast architecture with a clear decision protocol. Not one forecast, several. Not a single threshold, an ensemble. A protocol that specifies what combination of signals fires the trigger, with named decision authority for the release.

A grievance and accountability mechanism designed for the speed of disbursement. Anticipatory cash means people receive transfers before the disaster materialises. Some of them will not understand why. The communication and complaints architecture has to be ready before the funds move.

An impact verification framework built on counterfactuals. Standard PDM doesn't cut it for anticipatory action. The evaluation has to compare what happened to a plausible alternative where the trigger didn't fire. That requires comparison groups, agreed methodology, and donor acceptance of the inherent uncertainty.

An institutional learning loop. Every trigger, fired or unfired, generates evidence about the system. False positives, false negatives, lead-time accuracy, beneficiary outcomes — all of these feed back into the next iteration of the trigger. Building that loop is harder than building any individual component.

The COP30 Loss and Damage Fund and the Belém Adaptation Indicators have created the policy demand for anticipatory action evidence. DELTA Resilience and the G-DRSF have created the data architecture. Mobile money has created the disbursement rails. The forecasting science has matured. The case-study evidence is overwhelming.

What remains is institutional courage — donors willing to release funds before the disaster declares itself, governments willing to authorise pre-event transfers, implementing partners willing to be evaluated on counterfactual outcomes.

Every dollar of climate finance that arrives after the disaster is a dollar that could have prevented the disaster's worst consequences if it had arrived two weeks earlier. The data systems exist to make that timing possible. The question is whether we'll use them.