Aid — types + effectiveness
Bilateral, multilateral, NGO, emergency. Tied vs untied. Works when targeted, untied, long-term + accountable.
Official Development Assistance (ODA) is grants + low-interest loans from HICs to LICs for development. Global ODA in 2023 was ~$211bn (OECD).
Donor scale.
| Donor | Annual ODA | % of GNI |
|---|---|---|
| USA | ~$66bn | 0.24% |
| Germany | ~$36bn | 0.79% |
| EU institutions | ~$24bn | — |
| Japan | ~$20bn | 0.40% |
| UK | ~$19bn | 0.5% (cut from 0.7% in 2020) |
| France | ~$15bn | 0.47% |
| Sweden | ~$5bn | 0.90% (one of few above 0.7%) |
The 0.7% target (UN agreed 1970) is met only by Norway, Sweden, Denmark, Luxembourg, Germany. UK met it 2013-2020 then cut to 0.5% under fiscal pressure.
Types of aid.
| Type | What it is | Example |
|---|---|---|
| Bilateral | Direct government-to-government | UK FCDO funding Bangladesh health programmes |
| Multilateral | Channelled through institutions | USA contributions to World Bank IDA fund |
| NGO | Delivered by non-profits | Oxfam, BRAC, MSF, WaterAid programmes |
| Emergency / humanitarian | Short-term post-disaster | 3bn/yr |
| Tied | Recipient must buy donor goods | US food aid required to ship on US vessels |
| Untied | No purchase restrictions | OECD DAC promotes untying — more efficient |
Tied aid inflates costs by 15-30% (OECD estimate) because recipients pay marked-up prices to donor-country suppliers.
Does aid work?
Cases where aid works.
- Health: Gavi Vaccine Alliance has vaccinated ~1 billion children since 2000. Global Fund (AIDS/TB/malaria) cut African malaria mortality ~50% since 2000. Gates Foundation reduced polio cases from ~350,000/yr (1988) to <10 (2023).
- Education: aid-funded school construction + teacher training in Bangladesh, Rwanda, Ethiopia.
- NGO-driven scale: Bangladesh's BRAC (founded with foreign aid 1972) now reaches ~140 million people across 11 countries.
- Country transformations: Vietnam, South Korea, Botswana all benefited from sustained aid alongside reforms.
Cases where aid fails.
- Corrupt regimes: Mobutu's Zaire received ~5bn personal fortune.
- Tied aid distortions: aid forces recipient to buy donor goods at inflated prices, undermining local industry.
- Dependency: long-term aid without exit strategy can suppress domestic capacity.
- 'Dead Aid' critique: Dambisa Moyo (2009) argued $4tn+ aid to Africa since 1960 has not ended poverty + has fostered corruption.
What separates success from failure?
Aid works when:
- TARGETED at measurable outcomes (vaccination, schooling, water access).
- UNTIED + flexible.
- LONG-TERM + predictable.
- Channelled through ACCOUNTABLE local institutions or proven NGOs.
- Recipient has decent governance.
Aid fails when:
- TIED to donor exports.
- Goes to corrupt or unstable regimes.
- Bypasses local institutions.
- Used for donor political objectives, not recipient need.
A* synthesis. The 'does aid work?' debate is the wrong framing. The right question is 'what kind of aid, designed how, works in what context?' — and the answer is targeted, untied, long-term, accountable aid for health + education delivers; opportunistic, tied, regime-supporting aid does not.
- ODA ~$211bn/yr globally; UK 0.5% GNI; UN target 0.7%.
- Bilateral / multilateral / NGO / emergency.
- Tied aid inflates costs 15-30%.
- Successful aid: Gavi ~1bn children; polio cases 350k → <10; BRAC ~140m reach.
- Failed aid: Mobutu's Zaire $8.5bn → personal fortune.
- Design + context determine effectiveness.