Why the Green Climate Fund's simplified track speeds money to the ground but does not make it fairer, and what the Map of Maps of Governance learns from its first test outside India
Abstract
Does the governance of climate money behave like the governance of a ration card? The Map of Maps of Governance (Tailwinds 2026-02) proposed seven rules for how success and failure move through Indian institutions. This companion to Tailwinds 2026-10 takes the rules out of India for the first time and tests them, in a registered analysis, on the full portfolio of the Green Climate Fund (372 projects, frozen 12 October 2026). Three rules could be tested on the Fund's records; four were declared untestable before any result was seen. One rule holds. Islands: work moved out of the routine, high-discretion track succeeds, and projects on the Fund's Simplified Approval Process reach first disbursement 5.4 months sooner than regular proposals (8.8 against 14.1 months; 90 percent interval −7.3 to −2.9), in every run. Two rules fail. Downhill: delay does not fall harder on the most vulnerable countries (+0.6 months, −2.0 to 2.8). Broker: international intermediaries do not carry more of the money where vulnerability is higher; they carry about four-fifths of it everywhere (79.6 against 81.4 percent). The fast lane is faster, and it is not fairer: inside it, the most vulnerable third of countries waits 1.4 months longer than the least, against 0.2 months on the regular track. The rules travel in part. Where they fail, they fail for a reason the map can use: between countries, a fund's queue runs through its intermediaries, which serve every country alike. The tail that Downhill predicts lies inside countries, where the Fund's record cannot see.
Keywords
adaptation finance; Green Climate Fund; governance; propagation rules; intermediaries; Simplified Approval Process; Map of Maps of Governance; MOM-G; Tailwinds
Interactive map
Counted but Unknown: interactive map
Suggested citation
Ramprasad, V. 2026. The fast lane and the long chain: Why the Green Climate Fund's simplified track speeds money to the ground but does not make it fairer, and what the Map of Maps of Governance learns from its first test outside India. Tailwinds Working Paper 2026-11, Tailenders Analytics, v0.1, 31 October 2026. https://doi.org/10.5281/zenodo.24047096
Working papers share research in progress to invite comment; this version has not been peer reviewed. Comments are welcome through tailenders.org. © 2026 Tailenders Analytics. Licensed under CC BY 4.0: share and adapt with attribution.
A rule of governance that holds only where it was found is a description. A rule that travels is a theory. Tailwinds 2026-02 drew seven propagation rules from Indian governance: how a pension delay, a rejected claim or a reform moves through offices, brokers and households. This paper asks whether they hold for an institution that looks nothing like an Indian district office: a global fund that moves US$21.31 billion through accredited entities to 136 countries.
The answer is a scorecard, delivered as it fell. One rule holds, two fail, and four cannot be tested on the record the Fund publishes. Each verdict teaches the map something. The rule that holds names the Fund's real achievement. The rules that fail show where the action is not. The rules that cannot be tested show what the Fund does not record.
This paper makes three claims. First, Islands holds: the Fund's simplified track is a working island, faster in every run. Second, it is not a fairer island: speed does not close the gap for the most vulnerable countries. Third, Downhill and Broker fail between countries because the chain that matters is the one inside them. Reaching the most vulnerable countries is not the same as reaching the most vulnerable people, and the rules that predict how failure flows to the tail will be tested only when the record follows money to people.
Each rule was translated into a prediction about the Fund before the data were opened, and each was declared testable or untestable in the registration. The tests that could be run used the Fund's open portfolio, the INFORM Risk Index (Mid 2026) as the primary measure of country vulnerability, ND-GAIN (2023) as a sensitivity, and a third run restricted to single-country projects.
| Rule (Tailwinds 2026-02) | Prediction for the Fund | Result | Verdict |
|---|---|---|---|
| Islands | Simplified Approval Process reaches first disbursement sooner than regular proposals | 8.8 vs 14.1 months; −5.4 (−7.3 to −2.9) | Holds, in all three runs |
| Downhill | The most vulnerable third of countries waits longer for first disbursement | 13.3 vs 12.7 months; +0.6 (−2.0 to 2.8) | Fails, in all three runs |
| Broker | International intermediaries carry a larger share where vulnerability is higher | 79.6 vs 81.4 percent; −1.7 points (−11.2 to 8.0) | Fails, in all three runs |
| Exit cascade | Not tested | Untestable on the record | |
| Fear symmetry | Not tested | Untestable on the record | |
| Threshold | Not tested | Untestable on the record | |
| Neighbor | Not tested | Untestable on the record |
Delay is measured from Board approval to the first disbursement. Sixty-three approved projects have never disbursed and are excluded from the timing tests. A project's vulnerability is the mean of its countries'; thirds are cut across recipient countries. Intervals come from bootstrap resampling, clustered by country, at 90 percent.
The Fund's Governing Instrument promised "simplified processes" for small-scale activities (paragraph 53). It delivered. Projects approved through the Simplified Approval Process reach their first disbursement in a median of 8.8 months; regular funding proposals take 14.1. The gap holds under INFORM (−5.4 months, −7.3 to −2.9), under ND-GAIN (−5.4, −7.3 to −2.9) and in single-country projects (8.7 against 13.3 months; −4.6, −6.4 to −2.0). This is the rule as 2026-02 stated it: success holds where a task is converted out of the routine, high-discretion cell. A fund that standardizes small projects and caps their size moves money faster. The comparison does not hold project size or sector constant, and simplified projects are small by rule, so part of the gap may be size rather than process (Section 7).
Faster is not fairer. Inside the simplified track, the most vulnerable third of countries waits 1.4 months longer than the least vulnerable (2.7 months under ND-GAIN). Inside the regular track, the gap is 0.2 months (0.7 under ND-GAIN). These are point estimates within subgroups, not registered tests, and they are reported as a pattern to watch. An island lifts everyone on it. It does not, on its own, carry the tail.
Nor does the island keep a cleaner count, or a worse one. Of the 18 single-country projects whose direct and indirect beneficiaries add up to more than the national population (Tailwinds 2026-10), 8 came through the simplified track, which accounts for 62 of the 216 single-country projects with population data. That pattern disappears under the most generous reading: the 3 projects that still exceed the population are all regular proposals. The count problem belongs to the Fund's definition, on both tracks.
In Indian governance, delay flows to whoever has the least capacity to exit: the household without a contact, the claimant without a broker. The prediction for the Fund was that the most vulnerable countries would wait longest. They do not. The most vulnerable third waits a median of 13.3 months to first disbursement, the least vulnerable 12.7: a difference of +0.6 months (−2.0 to 2.8), indistinguishable from zero. Under ND-GAIN the sign reverses (−0.5, −2.9 to 1.7); in single-country projects it is −0.9 (−2.7 to 2.2).
Censoring does not hide it either. An exploratory count, outside the registration, shows that never-disbursed projects are not concentrated in the most vulnerable countries: 26 of 133 projects (19.5 percent) in the most vulnerable third under INFORM, against 19 of 86 (22.1 percent) in the least; under ND-GAIN, 13 of 118 (11.0 percent) against 22 of 113 (19.5 percent).
The failure says more about scale than about the rule: a country is not a household. A country does not queue at the Fund's door; an accredited entity does, and the large international entities that carry most of the money queue equally well for every country they serve. Downhill predicts where delay lands among people with unequal power to exit. Between countries, at the Fund's door, that inequality is absorbed by the intermediary. Inside countries, between the intermediary and the household, it is not, and that is where Downhill should be tested next.
The Broker rule says that intermediaries earn from complexity, and that discretion moves to whoever controls the chokepoint. The prediction for the Fund was that international intermediaries would carry a larger share of the money where vulnerability is higher and national capacity lower. They do not, because they carry about four-fifths of it everywhere: 79.6 percent of adaptation funding to the most vulnerable third of countries, 81.4 percent to the least (−1.7 points, −11.2 to 8.0). Under ND-GAIN, 75.6 against 80.3 percent; in single-country projects, 79.9 against 80.9.
The registered prediction failed, and the reason is more striking than the prediction. In this system brokerage does not track vulnerability; it is the system. Entities with direct access, the national and regional institutions the Fund created direct access to empower, carry 20.8 percent of all its funding. In Uganda, 80 percent of the Fund's adaptation money passes through international entities; in Bolivia, all of it. The long chain is the norm, and every link in it is a place where the record of who benefits can be lost.
Four rules were declared untestable before any data were opened. Each names a record the Fund does not keep, or does not keep in usable form.
| Rule | What it would predict for the Fund | The record it needs |
|---|---|---|
| Exit cascade | When capable entities leave a country's pipeline, quality falls and more leave | Accreditation lapses, withdrawals and pipeline exits by country, over time |
| Fear symmetry | Integrity pressure on staff and entities raises deferral, felt as delay | Disbursement timing before and after integrity investigations and suspensions |
| Threshold | Reform of the Fund's rules follows only shocks beyond what adjustment can absorb | Board decisions dated against disaster events and their losses |
| Neighbor | Direct access spreads between neighboring countries by imitation | Accreditation dates of national entities, with a border design |
The experience map offers a first trace, and it points at one of this paper's depth countries. The Fund's Independent Redress Mechanism lists 26 entries in its case register, complaints C0001 to C0024 and two requests for reconsideration, across a portfolio of 372 projects in 136 countries. Six of the 26 concern a single Ugandan wetlands project (FP034). Two of those six have reached compliance investigation, one of them opened on the Mechanism's own initiative (C0023); one ended in a problem-solving agreement between complainants from seven districts and Uganda's Ministry of Water and Environment, whose terms the parties kept confidential (C0010). Twenty-six entries do not measure harm; they suggest how far the people at the end of the chain are from the Fund's door, and how much louder they become when they find it.
The Map of Maps of Governance was drawn from Indian offices. Its first test outside India returns three lessons. First, Islands is the most portable rule: converting work out of the high-discretion track speeds it, in a district or in a global fund. Second, rules about exit and brokerage operate at the scale where unequal power meets the queue. Between countries, at a fund's door, that scale is the intermediary; the rules must be tested where the intermediary meets the household. Third, the map needs a fund layer: the accredited entity is a node with its own position in the four maps, and a fund's record should show the chain from Board to household, link by link.
That is the same conclusion Tailwinds 2026-10 reaches from the other side. The record that would test Downhill and Broker where they matter is the record that would tell the Fund who it reaches. Confidential reach verification builds it.
Islands would be overturned by evidence that the simplified track's speed comes from project size or sector rather than the process, once both are held constant. The failures of Downhill and Broker would be overturned by data from inside countries showing that delay and intermediation fall evenly on vulnerable and less vulnerable households, which would mean the rules do not travel at all, not that they were tested at the wrong scale.
The analysis uses the Fund's public project records, two published vulnerability indices and World Bank population. None contains the name of a person or household. Redress cases are counted from the public register and not described beyond what the register publishes. No entity or project is accused of misconduct.
The tests are between countries; the rules were built for relations between people and offices. Delay to first disbursement is one point on a long timeline, and it excludes 63 projects that have never disbursed. Vulnerability is measured at the level of the country by two indices that disagree. Intermediation is coded from accreditation type, not from who actually spends the money. The Islands comparison does not control for project size or sector. The within-track gaps and the never-disbursed counts are exploratory. Only one fund is examined.