Beyond ‘Does it work?’: three evidence priorities for scaling up anticipatory action

Beyond ‘Does it work?’: three evidence priorities for scaling up anticipatory action

Move on from trying to ‘sell’ anticipatory action and instead help shape how, where and to what ends it gets used. That is our main takeaway from the recent FAO-hosted event, ‘Leveraging evidence for effective anticipatory action’.

During the event, we heard from donors, national governments, practitioners and researchers, who are now asking: How can anticipatory action be designed to promote livelihood resilience, not just accelerate welfare gains? How should anticipatory action integrate with and complement longer-term resilience and social protection investments? And what is the expected return to anticipatory action once the total costs of system readiness are balanced against avoided humanitarian and recovery losses, and other benefits?

Participants also talked about the evidence they need to answer these questions – and how, collectively, we can all evolve our agendas to support this. From these discussions, we see three ways in which the evidence agenda for anticipatory action can respond to these and other emerging priorities.

1. From demonstrating impact to informing design decisions

Anticipatory action works – but not uniformly. Outcomes vary with context and programme design. In a recent synthesis, we found consistent patterns linking impacts to shock severity, cash transfer size and structure, and whether households can act meaningfully on early assistance. These relationships became apparent across studies ex post, but few evaluations were explicitly designed to test them.

The next generation of evidence on anticipatory action can start here. Research should pre-specify and test hypotheses about how and why design features shape outcomes. Priorities here include: transfer size and timing; targeting approaches; complementary (‘cash‑plus’) support; and whether outcomes are sustained over time.

FAO’s impact evaluation in Somalia, which tested lump sum and monthly tranche transfers, is a strong example. This is the type of evidence that policy-makers and implementers need — not to justify anticipatory action, but to deploy it optimally.

2. From stand-alone pilots to embedded anticipatory action

Much of the evidence base for anticipatory action has been built on stand-alone pilots. This is an appropriate starting point, and one that has generated valuable learning and proof points. But as anticipatory action is increasingly integrated into social protection systems, multi-year resilience portfolios and national architectures for disaster response, evidence must adapt accordingly.

Key questions on institutionalizing anticipatory action include: How does it perform when it shares delivery systems, budgets and decision authority with other tools? When does embedded anticipatory action represent good value for money relative to alternative investments in humanitarian response or resilience?  

Multiple case studies have described what such integration looks like, or at least what it should look like. More comparative analysis of the costs and benefits of embedded anticipatory action are needed. Without this evidence, we risk scaling up a model that was designed and evaluated in isolation into systems where its advantages – and its limits – look very different.

3. From household outcomes to system performance

Most evidence on anticipatory action captures household‑level outcomes. More of that is certainly needed (see priority 1), but policy and funding decisions hinge less on marginal household effects and more on how anticipatory action performs as part of a wider response system – and at what cost.

Donors need to know whether anticipatory action improves response speed, reduces caseloads and downstream needs, and enhances budget predictability. They're asking what the break-even activation rate looks like once the full costs and benefits of forecasting, readiness, pre-arranged financing and coordination are factored in. Answering these questions requires shifting the unit of analysis to produce comparative, system‑level cost evidence.

Much of the guidance and existing evidence on the costs of anticipatory action is focused on individual programmes. Again, more of that is needed. But without broader analysis, anticipatory action is often being judged on household‑level metrics alone, while the system‑level trade‑offs that ultimately govern scale‑up remain poorly understood.

Conclusions

As humanitarian budgets tighten, acting earlier is no longer optional; it is essential. But so too is the need for discipline. Scaling up anticipatory action without a clearer understanding of where it works best, at what cost and within which systems, risks overextending a model whose strengths and limits are still being mapped. The evidence agenda needs to respond to how anticipatory action is applied in practice: across diverse contexts, within complex systems, and alongside other response and resilience tools. Future funding for generating evidence should prioritize answering these questions, as they are the ones that shape outcomes most directly, cost‑effectiveness and the scalability of anticipatory action.