Four Futures for Foreign Aid

By the time you read this, the global aid architecture has lost roughly a quarter of its annual budget in a single fiscal year. Preliminary OECD data show that international development assistance from member states fell by about 23 percent between 2024 and 2025, with the United States contribution alone shrinking by nearly 57 percent — the steepest, fastest contraction of overseas development assistance the post-war system has recorded. For the first time, the OECD’s five largest donors — the United States, Germany, the United Kingdom, France, and Japan — all reduced their aid budgets in the same year.

This article argues that Monitoring, Evaluation and Learning (MEL) is not a bureaucratic afterthought in cybersecurity — it is the missing structural component that determines whether billions in public investment produce real, measurable security improvements. AI amplifies this urgency in two directions: adversaries are using generative AI, deepfakes, and machine learning to scale attacks at unprecedented speed, while defenders are deploying AI tools whose own effectiveness and risks must themselves be measured and governed. Drawing on global frameworks, regional comparisons across six continents, and practitioner experience, I set out a practical approach any government can adapt.

Among scenario planning techniques, the 2×2 matrix is the workhorse. It is fast enough to run in a one-day workshop, structured enough to discipline argument, and visual enough to communicate to a board. The method is simple in form: identify the two forces that combine highest impact on your strategy with highest uncertainty about how they will unfold; place each on an axis; describe the four resulting worlds.

Notice what the method explicitly avoids. It does not ask you to forecast a single number. It does not collapse the future into “best case / worst case,” which in practice usually means optimistic and slightly less optimistic. Instead, it forces you to live, briefly and seriously, in four different worlds — each internally coherent, each with its own implications for your theory of change.

For the aid sector specifically, the 2×2 has three virtues right now. First, it accommodates radical uncertainty: nobody honestly knows what the U.S. State Department’s restructured foreign assistance posture will look like in 2027, what European budgets will absorb in 2028, or whether the Gulf and Asian philanthropic surge — Gates Foundation’s pledged USD 200 billion spend-down by 2045, the projected USD 4.7 trillion in Asian billionaire wealth by 2026 — will translate into substitute development finance or something quite different. Second, it surfaces strategic options that are robust across multiple futures rather than optimised for one. Third, it builds shared mental models inside teams that have been cycling through panic and denial for over a year now.

Figure 1. The 2×2 places the two highest-impact, highest-uncertainty drivers on perpendicular axes. Each quadrant represents a coherent world worth describing in narrative form — not a probability, not a preference, just a plausible future to test strategy against.

The single most consequential decision in scenario work is axis selection. Get the axes wrong, and you produce four interesting but irrelevant stories. Get them right, and the matrix illuminates strategy for the next decade.

The classical guidance from Wack and Schwartz still holds: each axis must combine high impact on your focal question with genuine uncertainty about how it will resolve. A driver that is high-impact but predictable (climate change, demographic ageing in OECD donors) is what scenario practitioners call a predetermined element — it belongs in every scenario, not on an axis. A driver that is uncertain but low-impact is a distraction.

For a generic aid-cuts focal question — “How should our organisation position itself over the next 3–5 years?” — I propose the two axes below as a robust starting set. They are deliberately generic enough to apply to a multilateral, a bilateral donor, a large INGO, a national NGO, or a partner government planning ministry.

From SHARP CONTRACTION (current trajectory: ODA falls toward USD 151 billion by 2026, a near-29 percent drop from 2023, with Lancet projections of 9.4 to 14 million additional preventable deaths by 2030) to PARTIAL REBOUND (the bipartisan U.S. FY2026 appropriation of USD 50 billion, philanthropic surge from Gates and others, growing Gulf and Asian giving, and renewed European multi-year frameworks substantially restore — though never fully replace — pre-2024 levels).

From RECENTRALISED (donors channel through their own bureaucracies and a small number of trusted INGO primes; the State Department absorbs USAID functions; FCDO consolidates portfolios) to DEVOLVED & LOCALISED (funding flows directly to country systems, regional pooled funds, and national NGOs in line with — and well beyond — Grand Bargain commitments; INGOs reposition as networked intermediaries rather than primary deliverers).

Crossing these two axes generates the matrix shown in Figure 2. Each quadrant is named, dated to a notional 2029, and described with a short narrative, leading indicators to watch, and the key strategic implication.

Figure 2. The four futures, generated by crossing aid volume (vertical) with locus of delivery (horizontal). Names matter: a memorable noun phrase per quadrant is what makes the scenario stick in a board conversation six months later..

The point of running the matrix is not to bet on one quadrant. It is to identify what scenario practitioners call robust strategies — moves that pay off in multiple futures and are unlikely to fail in any. After workshopping versions of this matrix with country teams across Southeast Asia, Sub-Saharan Africa, and Latin America over the past year, four such strategies recur with striking consistency.

One. Diversify the funder base aggressively, with explicit non-OECD targets. Asian, Gulf, and South-South philanthropic and institutional flows are the only categories growing in absolute terms. Even in the Restored Status Quo scenario, dependence on a single bilateral donor is now a clear single-point-of-failure risk.

Two. Build and document genuine local partnerships now, not when a donor demands it. In every scenario except the most recentralising one, the organisations that prosper are those with a credible roster of national partners they have invested in over years. The “starvation cycle” critique — donors funding programmes but not the indirect costs that keep local NGOs solvent — must end on the implementer side first.

Three. Invest in evaluation, learning, and evidence as a competitive moat. In a world of contracting budgets, donors will increasingly fund what is demonstrably working. Adaptive management, real-time monitoring, and credible impact evidence are no longer nice-to-haves; they are the licence to operate.

Four. Plan exits and transitions from project inception. The 2025 cuts taught the sector a brutal lesson: the absence of pre-planned transitions turned every termination into a humanitarian event of its own. As The New Humanitarian observed in early 2026, “humanitarians are experts at surge and scale-up. They’re not so great at the dismount.” Exit planning, social-protection-system handovers, and “do no harm” closure protocols are now baseline competence.

The full matrix below is intended as a one-day workshop template. It is methodologically conservative — closer to Schwartz’s Art of the Long View formulation than to more elaborate French prospective traditions — because the binding constraint in most aid organisations right now is time, not analytical sophistication.

For donor agencies and foundations, the same template adapts with one modification. The focal question shifts from “How should we position ourselves?” to “What portfolio shape is most likely to deliver against our mandate across all four futures?” — a question that turns scenario planning into portfolio design.

Three suggestions follow naturally. First, allocate a meaningful share of the portfolio — practitioners I have worked with use figures between 10 and 25 percent — to instruments designed for the Sovereign Reset and Lean Handover quadrants: direct local funding, country-based pooled funds, and core support to national civil society networks. These instruments are still under-resourced relative to where the system is plausibly heading.

Second, build explicit transition clauses into every multi-year agreement. The 2025 experience demonstrated that abrupt termination is itself a humanitarian harm. Standard exit protocols, with funded handover periods and contingent recipient governance arrangements, should now be a default contract clause, not an exception negotiated under crisis.

Third, invest in shared infrastructure that is robust across all four scenarios — joint evaluation systems, common data standards, integrated MEL platforms. The fragmented evaluation architecture of the past two decades is a luxury the sector can no longer afford. The work my colleagues and I have done on the FRAME methodology for responsible AI in monitoring and evaluation is one entry point here, but the broader principle is straightforward: in a contracting sector, public goods become more, not less, valuable.