The word "governance" is doing real work here. This is not an organisation chart; it is a set of overlapping authorities with no single decision-maker and no appeal body.

Money is the fastest-moving of the three and the one practitioners feel first. A donor government's domestic politics can close a country programme in a single budget cycle.

Nothing in this chain is necessarily waste — audit, evaluation and compliance are real costs and are often what donors are legally required to fund. But every link is also a place where a decision gets made about what the money is for.

The phrase "India received $X billion in aid" almost always includes concessional loans. Those are debts, and reporting them as gifts overstates the transfer and understates the obligation.

The ordering is the point and it is stable across sources: a country's own tax base dwarfs everything donors provide. Aid is at most catalytic; it has never been the main event.

Knowing the architecture is not academic background. It tells you which requirements are genuinely non-negotiable — a statutory audit rule — and which are habit that a programme officer can waive.

Both institutions were designed by and for a world of 44 mostly Northern states — most of Africa and much of Asia was still colonised and had no seat at the table.

The mismatch between the third row's two halves — equal voice in the General Assembly, weighted voice where the money is — is the structural fact behind sixty years of reform demands.

The last row is not history. Several donors have merged their development agencies into their foreign ministries in recent years, which is a decision about whose interests aid serves.

The term was coined by John Williamson in 1989 as a description of what Washington-based institutions then agreed on. He later objected that it had been stretched into a label for market fundamentalism generally.

Each correction absorbed the previous critique rather than replacing the institutions that generated it. The Bank and the Fund made the turn to poverty; they were not displaced by it.

Nothing in this list was repealed. The architecture accumulates: each era adds institutions and rules on top of the last, which is why it is complex out of proportion to any single purpose.

The distinction matters practically. WHO and the ILO are specialised agencies with their own constitutions; UNICEF and UNDP are funds and programmes under the General Assembly. They are not interchangeable and do not answer to one another.

A resolution of the General Assembly is a statement of what the world says it wants. It creates political expectation and no legal obligation, which is a real form of power and a limited one.

The overhead row explains a chronic complaint: heavily earmarked agencies subsidise project delivery from a shrinking core, which weakens exactly the functions donors then criticise as inefficient.

The country framework agreed with the government is a public document and worth reading before any approach. It states what the UN system has committed to for the next several years, which is the shape of the funding.

The norm-setting record is the most durable achievement and the least visible. Conventions on the rights of the child, on discrimination against women, on disability rights are all now cited in domestic litigation.

The last two are the price of the first strength. An organisation that includes every state can only move at the speed of agreement, and cannot compel a member without its consent.

The last row is the one that explains the scale difference. The World Bank borrows on capital markets against its shareholders' backing, so a dollar of paid-in capital supports many dollars of lending.

The Fund's influence exceeds its lending because its assessment is a signal to every other creditor. A country off-track with the IMF finds bilateral and market finance harder to obtain — the "seal of approval" effect.

The NDB's equal founding shareholding is a deliberate contrast with the Bretton Woods model — it was designed to demonstrate that a development bank need not weight votes by wealth.

Because it is unwritten, it cannot be repealed — only outvoted. That is why every reform push targets voting shares rather than the convention itself.

The defence is that a lender may reasonably attach terms, and that conditions can strengthen a reforming finance minister's hand domestically. The critique is that unelected outsiders end up setting policy a electorate did not choose.

Quota reform at the IMF has moved in small increments and repeatedly stalled, because any realignment requires the states that would lose share to vote for losing it.

Recipient governments generally prefer multilateral channels for the transaction-cost reason alone: a country dealing with thirty bilateral donors is running thirty reporting regimes.

This is why "global aid fell this year" is usually a statement about DAC members only. Flows from non-DAC providers are real, substantial, and outside the series.

Some items count as ODA without leaving the donor country — in-donor refugee costs in the first year, and some scholarship and administrative costs. In several donor budgets these have been a sizeable share.

The grant element is calculated by discounting the future repayments, so it depends on the assumed discount rate. The DAC changed that methodology in 2018, which changed the reported totals without changing any money.

The DAC recommendation on untying aid to least-developed countries dates from 2001 and coverage has widened since, but tying by other means — technical assistance staffed from the donor country, for instance — is harder to observe.

These six questions will resolve most apparent contradictions between two published aid figures for the same country and year.

The final row is the structural weakness. The goals were agreed in September 2015 and the financing framework in July 2015, as separate negotiations, and the second does not fund the first.

The MDG-era lesson behind the pledge is concrete: a national target can be met by improving outcomes for those easiest to reach, while the poorest tenth stagnates or falls further behind.

"Voluntary" is doing a lot of work. A VNR is written by the government reporting on itself, with no verification step and no consequence for omission.

Indicators are graded by methodological maturity — check the tier before citing one.

The interactions run both ways and the literature on them is substantial. Treating the goals as a menu — picking the three your programme addresses — is exactly the reading the framework was designed to prevent.

NITI Aayog's SDG India Index scores states and union territories against a national indicator set. It is genuinely useful and it is a self-assessment, built on indicators India selected.

The Addis Ababa Action Agenda's central move was to put domestic resource mobilisation first and treat aid as one source among several — a reframing that donors welcomed for obvious reasons.

The accountability argument is the one most often skipped and the most consequential: a government funded by its own taxpayers has to answer to them in a way that an aid-funded one does not.

The last row is not illicit at all, which is why it is often left out of the headline figures and is arguably the largest and most tractable item — a government can repeal its own exemptions.

The last row is the structural change of the past two decades. Debt used to be owed mainly to a small club of official creditors who could meet and agree; it is now spread across bondholders and non-Paris-Club lenders.

Note that three of the four are outside the borrowing government's control. Debt distress is frequently narrated as imprudence, and is often a shock arriving on top of an exposed balance sheet.

The critique is not that blending never works. It is that the leverage ratios used to justify it have been optimistic, and that the instrument gravitates to places and sectors that already attract capital.

The Addis conclusion follows from the table: no single source funds the agenda, and each is suited to different things. Aid used for recurrent salaries is aid creating a cliff.

Plurality genuinely increases recipient choice, which is a real gain in bargaining power for a government that previously faced a single donor consensus.

The scale is real: in some areas of global health, private philanthropic funding is comparable to that of large donor governments, and it can be deployed faster and with fewer domestic political constraints.

The framing has real substance — the absence of policy conditionality is a genuine difference, and recipients say so — and it is also a positioning strategy in a competitive market.

The "debt trap" framing is contested by researchers who have examined the loan portfolios: distress cases exist, and the evidence for a deliberate strategy of seizing assets is weak.

Researchers have built estimates from contract-level data, and those estimates are the best available while being methodologically contested. Cite them as estimates.

PPP contingent liabilities are the underappreciated item: guarantees given to private partners sit off-budget until they are called, at which point they are debt.

The critique is largely internal to the sector and has been made most forcefully by people working inside it — which is why it has produced concrete commitments rather than only argument.

For a strong finance ministry with capacity, choice is an unambiguous gain. For a weak one, thirty uncoordinated funders is a governance problem in itself.

The classic illustration is a health ministry writing hundreds of separate donor reports a year, with senior staff time going to compliance rather than to health.

Busan mattered because it tried to include providers outside the DAC. They participated on terms that made the commitments voluntary for them, which is both why they came and why the regime weakened.

The tension with conditionality is not resolvable by better intentions. A lender attaching conditions and a borrower setting its own priorities are, at the limit, the same decision made by two parties.

The risk row is the sharpest criticism made by national organisations: they are asked to take on delivery in difficult settings while liability and reputational protection stay with the international partner.

Annual independent reporting has been unusually frank about slow progress.

The distinction from localisation is real. Localisation is largely about the flow of funds and authority; this is about the categories the sector uses to describe people, which survive a change of funding route.

The capacity argument deserves scrutiny wherever it appears. It is sometimes true and it is unfalsifiable as usually stated, which makes it an effective way of deferring a transfer indefinitely.

The three positions are less contradictory than they look, because they are about different things: specific interventions, political incentives, and aggregate growth.

The last row is concrete and under-discussed: donor-funded organisations pay above local public-sector rates, which draws skilled staff out of ministries that then lack capacity.

Dead Aid (2009) was aimed specifically at government-to-government aid in Africa, not at humanitarian or health assistance — a distinction frequently lost when the book is cited.

The cash-transfer evidence is the sharpest rebuke to assumed preferences: given money without conditions, people spend it broadly as they said they would, and not on the things paternalist design assumed.

The last row is where conditionality becomes acutely visible: fuel and food subsidy removal has repeatedly produced unrest, and the government carries the consequence of a condition it did not choose.

The proposal-language point is practical rather than abstract: a funding round that requires a theory of change in a particular format selects for organisations that have someone who can write one.

The asymmetry is not an oversight. Upward accountability has a legal instrument behind it, and downward accountability generally does not, which is why exhortation has not closed the gap.

The reason to hold both is practical. A practitioner who believes only the critique cannot work inside the system; one who believes only the defence cannot see what needs changing.

The 1991 crisis is the pivot in the domestic memory of this relationship, and it shapes the framing of Indian development cooperation today — conditionality is remembered from the receiving end.

Lines of credit are the dominant instrument by value, and they are concessional loans — so most Indian development cooperation, like most global development finance, is debt rather than grant.

The sourcing point is not a hidden criticism — it is how lines of credit are structured, and it is a form of tied aid by the standard definition, whatever the framing.

The African Union's admission is the concrete outcome and a real change in the composition of the forum. The rest is agenda-setting, which matters and is not the same as delivery.

Each demand runs into the same structural fact: the bodies that would have to approve the change are governed by the states that would lose from it.

Naming these tensions is not a criticism unique to India. Every provider of development cooperation carries a version of them, and the exercise is worth doing for whichever country you work in.

The single most useful habit from this course is small: whenever you meet a development figure, ask who counted it, under whose definition, and whether it is a gift or a loan.