Anticipatory Action and the Data Evidence: Why Waiting Is No Longer Caution

*The 2023–2024 El Niño drought response in southern Africa is the cleanest piece of evidence the anticipatory action field has produced. The lesson is not theory. It is what happened the last time the system tried to act early, and it worked.*

A few years ago I was in my office attempting a pilot design for drought anticipatory action triggers for a humanitarian response in Afghanistan. The forecast layers were good. The food-security, agricultural and livelihood needs and gaps data was current. The pre-positioning priorities were well understood by the actors. What was missing was the institutional permission to release the funding against a forecast rather than against an observed impact. The conversation that day was about evidence. The older argument, anchored in widely cited figures like the Global Commission on Adaptation's "Adapt Now" report finding that US$1.8 trillion invested in five resilience areas could generate roughly US$7.1 trillion in net benefits, was credible but largely anecdotal in the way the donor side experienced it. There was no recent, attribution-clean case study showing that an anticipatory disbursement, at scale, actually outperformed reactive disbursement against the same hazard. So the conversation moved sideways. Some agencies restructured existing financing to do what preparedness they could within the rules they already had. Others quietly relabelled already-planned activities as "anticipatory," which many practitioners doubted at the time. Around the trigger discussion itself, there was real skepticism that the cost-benefit case would hold once the dollars were on the table. The decision on a fresh anticipatory release was deferred. The drought arrived. The response that did happen happened later, at higher cost, and reached fewer households than an early activation would have.

Two years on, the evidence the field then lacked is on the table. The argument for acting early is no longer an appeal to logic. It is a citation. And with a strong El Niño now developing through 2026, per WMO, the question is whether the data has finally outpaced the institutional caution that used to be its only counterweight.

The cleanest case study the field has is the 2023–2024 southern Africa El Niño drought response, evaluated by CGIAR. Every dollar of anticipatory investment delivered early carried the humanitarian impact of US$1.30 delivered later. A 30 percent efficiency gain on the identical dollar.

Translate that operationally. For a cash transfer coordinator, that means the same transfer envelope reaches further. For a food-security programme, that means the same food basket prevents more acute malnutrition. For a multi-cluster response, that means the same overall financing covers more households at the same level of protection. Thirty percent is not a marginal improvement. In an environment of stagnant humanitarian funding and rising needs, it is the difference between meeting the appeal and falling short.

The southern Africa finding is not an outlier. It sits inside a convergent body of evidence. Modelling on the economics of resilience and early action across Ethiopia, Kenya and Somalia finds US$2.3 to US$3.3 in net benefits per US$1 invested. Multi-country analysis from FAO puts avoided losses at up to US$3 for every US$1 spent ahead of drought. The FAO/OCHA/WFP evidence base on anticipatory action consolidates the case across geographies. Tufts and colleagues have separately mapped the landscape of anticipatory action for health, with parallel findings on the value of acting before the disease pathway opens.

Three things are worth noting about this evidence. It is recent enough that the institutional and operational conditions still hold. It is methodologically diverse enough that the result is not an artefact of one evaluation framework. And it is geographically distributed enough that it cannot be dismissed as a single-context success. This is not theory. This is what the system did when it tried.

The ROI figures are the easy part of the argument. They are quantifiable, defensible and powerful in budget conversations. But they sit on top of a deeper set of outcomes that no dollar figure can fully represent, and those outcomes are the actual case for acting early.

The 2023–2024 southern Africa anticipatory response reached almost 2 million people before the worst drought impacts hit. That number, read in operational detail, describes households that did not have to sell their livestock at distress prices. Children who did not slip into acute malnutrition because the cash transfer reached the household before food prices spiked. Families who did not split because the income source held. Communities that did not fragment into migration because the productive assets the local economy depends on were preserved.

Early cash means food bought before prices spike. Early water-system maintenance and pre-positioning means disease outbreaks contained before they spread. Early livestock support means herds not decimated. Early protection programming means protection risks that never become protection crises. None of those outcomes turn up clearly in a cost-benefit ratio. All of them are the actual purpose of the response.

The quantitative case is necessary because it gets the conversation through the budget meeting. The qualitative reality is suffering prevented at a scale a dollar figure cannot hold. We should make both arguments, in that order, and resist the temptation to let the easier metric crowd out the harder one.

Beyond the immediate efficiency and humanitarian arguments lies the most under-appreciated case for early action. Anticipatory response preserves the foundations of resilience that future shocks will land on.

The mechanism is straightforward and visible across every protracted-crisis context I have worked in. When a household is forced to liquidate productive assets to survive a shock, the household enters the next shock weaker. When farmers abandon fields, the social capital and local agricultural knowledge that the next season depends on is partly lost. When children are pulled from school during a drought, their lifetime earnings trajectory bends in ways the household will not recover from. When families fragment under migration pressure, the support networks that absorb the next shock are no longer in place.

Each shock that hits a household at depleted capacity erodes the base from which it absorbs the next. The trajectory is downward. Anticipatory action interrupts that trajectory. By preserving livestock, productive assets, family structures, school enrolment and local knowledge, it allows households to enter the next shock with their base intact. Resilience compounds. So does its erosion. The choice between early and late action is, in part, a choice between which of those compounding effects we put under way.

This logic is decisive in the Horn of Africa contexts I know best, in Ethiopia and Somalia, where conflict is ongoing, the climate trajectory is worsening, and shocks land before recovery from the previous one has completed. In those settings, the difference between acting early and acting late is the difference between communities that absorb shocks repeatedly and communities that fragment permanently. That is not a technical preference. It is a policy choice about which institutions and which households still exist a decade from now.

The standard objection to acting on a forecast used to be that the forecast was not reliable enough, the financing was not flexible enough, or the operational architecture was not in place to translate a trigger into a response at scale. None of those objections still survives the current state of the field.

Forecasting has matured beyond the threshold the early debate set. Multi-model ensembles produced by national meteorological agencies and regional climate services are now reliable enough to trigger action with operational confidence on a range of slow-onset hazards. Impact-based forecasting, the discipline of moving from "drought is coming" to "drought will cut forage by 40 percent in this pastoral zone" or "flooding will block access to these health facilities for these days," has moved from research into operational use. It enables specificity. Specificity enables precise targeting. Targeting is the difference between cash that reaches the right household and cash that does not.

The financing architecture is operational. CERF, IFRC DREF, the Start Fund and Start Ready, FAO SFERA and the WFP Anticipatory Action Fund collectively form a financing layer that did not exist a decade ago. Pre-arranged agreements can disburse in days, and anticipatory action frameworks are now operational across dozens of countries through these agencies and a broader community of practice that has matured significantly over the last five years.

The data and delivery infrastructure is in place. Mobile money allows cash to reach a household in minutes once the trigger is initiated. Pre-positioned supplier networks shorten the delivery chain. SMS-based early warning bridges the last mile to affected populations. Satellite monitoring produces the impact data the trigger architecture depends on. And, as I have argued in DELTA Resilience as the data backbone for anticipatory action, the loss-data infrastructure that determines the threshold at which anticipatory action should be initiated is finally being treated as essential infrastructure rather than optional reporting.

There is one further enabler that deserves naming. Locally-led anticipatory action, with community knowledge recognised as essential rather than peripheral, is finally beginning to receive its own funding lines. The evidence base from the last two cycles is unambiguous that the responses that worked best were the ones that local institutions and local actors led. The architecture is starting to align with that finding.

In June 2026, with a strong El Niño developing, per WMO, three facts stack in favour of acting now.

The recent evidence on anticipatory investment is unambiguous. The 2023–2024 southern Africa response delivered a 30 percent efficiency gain. Horn of Africa modelling delivered US$2.3 to US$3.3 in net benefit per dollar. Multi-country analysis delivered up to US$3 in avoided losses per dollar. The pattern across geographies and methodologies is consistent enough that the burden of proof has shifted onto the case for delay.

Acting early preserves the foundations of resilience in contexts that have lost the slack to absorb compounding shocks. The downstream effects of a single avoided distress sale, or a single livestock herd preserved, or a single school year not abandoned, propagate forward into the household's capacity to survive the next event.

We have the forecasting, the financing, the data infrastructure and the operational frameworks to act at scale, now. The activations already under way across Central America, East Africa and the Sahel, through FAO, WFP, IFRC and partner agencies, are the visible evidence that the architecture works. The constraint is no longer capability.

The only barriers left are the speed of decision-making and the flexibility of funding. Both are choices, made by institutions, on a timeline they can change.

Waiting for impacts to materialise before acting is no longer caution. The evidence is too strong. The operational architecture is too mature. The cost of delay, in fiscal terms and in human terms, is too well-documented to be defensible by a precautionary framing. Inaction in the face of this much converging evidence is a choice the field can no longer claim it did not have.

I designed drought triggers in an office where the case for acting early was a strong intuition. It is now a citation. The data has caught up to the argument. The argument is the policy. The policy is the trigger we pull.

*Tags: Anticipatory Action · Data & Evidence · Climate Finance*