Gravitywell ResearchInstitutional Investment Intelligence

Global Capital Markets / Development & Climate Finance

The $7 Trillion
Paradox.

Why committed capital is not deploying into emerging markets — and the capital-markets re-plumbing that could finally move blended finance from billions to trillions.

~$4tn

Annual SDG
financing gap

$18bn

Annual blended
market · 123 deals

7.6x

Peak leverage
on $1bn+ deals

$249bn

Cumulative to date
· 1,350 deals

GWR-2026-BF-001 · June 2026 · gravitywell.io Institutional · For Professional Investors

Gravitywell Research — The $7 Trillion ParadoxContents

Inside This Report

CONTENTSNavigating the Deployment Gap

Front Matter

00Executive Summary03

The paradox quantified · six findings · the decision-ready conclusion

—Investor Quick Read06

Five conviction calls · five material risks (severity-rated) · the bottom line

Part I · The Deployment Gap

01Anatomy of the Gap07

Committed vs. deployed · leverage that scales with size · regional & sectoral flows

02The Barriers10

Pipeline · Basel/Solvency/fiduciary · FX & macro · ESG · politics · China

Part II · What Works & What It Costs

03Instrument Effectiveness14

Effectiveness Index · where instruments fail · cost-per-mobilized-dollar · SDG Loan Fund

04Returns, Losses & the Risk Re-Rating18

Realized IRRs by tranche/region · GEMs granular data · failure cases

Part III · The Re-Plumbing

05The Capital-Markets Re-Plumbing22

O2D · securitization mechanics · structure registry · from pilot to platform

Part IV · Scrutiny, Evidence & Outlook

06Counterfactual & Critique26

Additionality · crowding-out evidence · the borrower's-eye view

07Transaction Registry & Case Studies28

Named deals across climate, energy, infra, health, inclusion, agri

08Outlook & Investment Implications31

Scenarios & flip-triggers · reform ledger · contingent-liability accounting · allocator playbook

Part V · The Private-Capital Desk

09What You Can Actually Buy35

What's raising, what it costs, how you exit — and the case against. Dated May 2026.

Appendices

—Methodology & Source Register45

Measurement regimes · contested figures · full citations

—Glossary & Acronyms47

Key blended-finance terms and institutional abbreviations, defined

The Thesis

The $7tn paradox is not a shortage of money or risk appetite. It is a shortage of the standardized, rated, liquid, fiduciary-compliant instruments that institutional capital is permitted and structured to buy. Capital is mis-wrapped, not missing.

Why It Matters Now

Blended finance that genuinely mobilizes private capital runs at ~$18bn/yr against a ~$4tn SDG gap — a 1–2% solution. Yet leverage scales sharply with deal size (7.6x on $1bn+ deals), and a securitization architecture is being built in real time. This report maps where it works — and where it doesn't.

Editorial Standard

Every figure is sourced and dated to primary documents (numbered citations, §Methodology). Where measurement is contested across the OECD, MDB Joint Report and Convergence regimes, divergences are disclosed, not reconciled. Forward-looking and self-reported figures are flagged. This edition adds realized return/loss data, a counterfactual critique, failure cases, and a borrower-side view.

Gravitywell Research · Confidential02 / 47

00 — Executive SummaryThe $7 Trillion Paradox

Executive Summary

00A 1–2% Solution to a Trillion-Dollar Problem

The global development-finance system is sitting on multi-trillion-dollar pools of committed and pledged capital — from DFIs, MDBs, sovereign and climate funds, foundations, pension funds and impact investors — yet the capital that actually reaches emerging-market and sustainable-development projects arrives in a comparative trickle. Blended finance that genuinely mobilizes private money runs at roughly $18 billion a year; MDB private mobilization adds on the order of $70 billion. Set against an SDG financing gap near $4 trillion and an annual climate-finance shortfall above $5 trillion — the gap between the ~$7.4tn a year needed to 2030 and the ~$2tn now flowing — the system is delivering a one-to-two-percent answer. The title's "$7 trillion" is the order-of-magnitude combined annual need this capital must meet — roughly $4tn for the SDGs plus the ~$3tn EMDE share of the climate transition; the paradox is that committed capital dwarfs it yet barely deploys.

~$4tn

Annual SDG financing gap, developing economies

~$18bn

2024 blended-finance market (Convergence 2025)

~$70bn

Annual MDB private mobilization (OECD/Joint Report)

1.8x

Private capital per concessional $ — flat since 2018

The Deployment Gap, VisualizedFIG. 0.1

Annual need versus annual mobilized flows. Bars are not additive across measurement regimes; see §Methodology.

SOURCE: UN FSDR 2026 / UNCTAD WIR 2025 (~$4tn SDG gap); CPI Global Landscape of Climate Finance 2025 (~$7.4tn/yr need to 2030 vs ~$2tn flows); Convergence State of Blended Finance 2025 (2024 flows); OECD / MDB Joint Report 2023. Bars illustrative, non-additive.

The core efficiency metric — the leverage ratio, or private dollars catalyzed per concessional dollar — has been frozen near 1.8x since 2018. The system has optimized for deploying its own balance sheet rather than for crowding in third-party money. The constraint is structural, not monetary.

Gravitywell Research · Confidential03 / 47

00 — Executive SummarySix Findings

Executive Summary — Continued

00 · iiSix Findings That Reframe the Debate

Finding 01 · The Binding Constraint

It is a pipeline problem, not a capital problem.

The scarce resource is bankable, investment-ready projects — not allocator appetite. OECD's 2025 guidance concedes the field stayed a "cottage industry" of bespoke deals. Origination capacity, not money, is the rate-limiting step.5

Finding 02 · Leverage Scales With Size

The ratio isn't frozen — it's fragmented.

Average blended leverage is a weak ~1.8–4x, but $1bn+ deals reach 7.6x, and $500m+ climate deals with MDBs present hit 5.9x. Small bespoke tickets, not the model, are the problem. Aggregation is the fix.1,2

Finding 03 · Regulation Rejects Form

Basel, Solvency II and fiduciary rules penalize the wrapper.

EMDE assets are frequently sound but arrive unrated, illiquid and bespoke. GEMs data now show private default of 3.54% — comparable to advanced-economy sub-investment-grade firms — against far higher perceived risk.4,6

Finding 04 · Returns Are Real but Modest

This is a fixed-income-like asset class, not a venture one.

Diversified EM private-credit/impact strategies cluster at ~6–9% net; senior blended tranches yield less. Investors expecting equity-like returns will be disappointed; those seeking rated, de-risked yield will not.9,10

Finding 05 · Information Is the Cheapest Trillion

Disclosure beats another summit pledge.

The Center for Global Development sizes full GEMs disclosure as a $600–800bn "data dividend"6 — capital unlocked by re-pricing perceived vs. realized risk. GEMS 2.0, launching as a standalone entity, has unanimous MDB support, and at least one rating agency has signaled openness to revisiting its methodology.21,22

Finding 06 · The Additionality Caveat

Some "mobilization" isn't additional.

Rigorous study rejects complete crowding-out but finds no evidence of substantial additionality on average.15 A share of reported flows likely subsidizes deals that would have closed anyway. Capital-markets structures help by making the catalytic layer explicit and priced.

"Billions become trillions not by forcing more money into a broken pipe — but by replacing the pipe with plumbing the buy-side already knows how to use." Gravitywell Research — Editorial Thesis

Gravitywell Research · Confidential04 / 47

00 — Executive SummaryThe Decision-Ready Conclusion

Executive Summary — Conclusion

00 · iiiWhat To Do With This

Private Allocators

Buy the wrapper, not the project.

The near-term opportunity is the rated senior and mezzanine tranches of MDB originate-to-distribute programs (MCPP, CLOs, SRT), not direct single-project EMDE risk. Underwrite to a ~6–9% blended-yield profile, not equity returns.

SWFs, PE & Infra

Fill the equity gap.

Target mezzanine, anchor and catalytic-equity roles, with the ~$215bn annual clean-energy equity gap in EMDEs as the headline opening — the slice senior debt and first-loss cannot fill.

Governments & DFIs

Standardize and disclose.

The highest-leverage public action is information and standardization — full GEMs disclosure plus harmonized documentation — and re-engineering internal incentives to count mobilization, not disbursement.

The One-Line Conclusion

The $7 trillion is not missing — it is mis-wrapped. Replace the bespoke, unrated, illiquid pipe with rated, tranched, standardized capital-markets plumbing; supply the catalytic first-loss that makes senior tranches safe; and publish the data that proves EMDE risk is lower than priced. Do those three things and billions become trillions. Do none, and the paradox persists.

New in This Consolidated Editionv2.0

§04 Returns, Losses & Re-Rating — realized IRRs by tranche/region, granular GEMs default-recovery data, and documented failure cases.
§06 Counterfactual & Critique — the additionality and crowding-out evidence, plus a borrower's-eye view of cost and speed.

Cost-per-mobilized-dollar instrument ranking for policymakers (§03), and contingent-liability accounting for guarantees (§08).
China, macro & FX added to the barriers (§02); all headline figures refreshed to 2025 primary sources with numbered citations.

Gravitywell Research · Confidential05 / 47

Investor Quick ReadFor Time-Constrained Readers

The 90-Second Version

QUICK READFive Calls, Five Risks

If you read nothing else: the constraint is structural, not monetary. The asymmetric opportunity sits in the credit-enhancement layer; the binding risk is the supply of catalytic first-loss capital. Conviction levels and material risks below, each cross-referenced to its evidence chapter.

"Capital is mis-wrapped, not missing. The winning trade is to own the standardized senior paper the re-plumbing creates — not the bespoke blend beneath it." — CORE THESIS

Five Conviction CallsCONVICTION

1 · Back the enhancement layerHigh

Guarantees and first-loss capital are the highest-leverage, most undersupplied instruments — the asymmetric exposure. → §3.1

2 · Own senior tranches, not the blendHigh

The institutional trade is rated, standardized senior securitized paper — not bespoke single deals. → §5.1

3 · Treat the re-rating as an optionMedium

A GEMs-disclosure re-rating is a real option with large upside, not a base case. Size it as such. → §4.2

4 · Watch the catalytic-capital tapHigh

First-loss supply is the single swing variable across every scenario. It gates everything downstream. → §8.2

5 · Underwrite FX, not just creditHigh

Local-revenue/hard-currency mismatch converts performing projects into defaults. Price it explicitly. → §2.4

Five Material RisksSEVERITY

■

Catalytic-capital scarcity

First-loss supply is structurally short; caps system leverage near 1.8x.

High
Structural

■

Re-rating never arrives

Rating agencies & regulators may not respond to GEMs disclosure.

High
Persistent

■

FX & macro tightening

Higher-for-longer USD widens spreads and deepens local-currency stress.

High
Active

▪

ODA cuts / retrenchment

Donor withdrawal starves grants, TA and the first-loss pool.

Medium
Rising

◦

Additionality / greenwashing

Crowding-out critique & weak KPIs erode credibility, not capital.

Monitored
Reputational

■High severity▪Medium◦Monitored

The Bottom Line

A ~$18bn/yr market against a ~$4tn gap is a 1–2% answer — but the architecture that could close it is being built now. Position in the credit-enhancement and senior-securitization layer, treat the regulatory re-rating as upside optionality, and monitor catalytic-capital supply as the leading indicator. Full conviction calls and the allocator playbook in §8.

Gravitywell Research · Confidential06 / 47

Part I · The Deployment Gap

Section 01

Anatomy of
the Gap

Committed capital versus deployed capital. The financing gaps quantified. And the finding that reframes the paradox — leverage that scales sharply with deal size.

Reading

Pages 07 — 09

Key Figure

7.6x on $1bn+

What You'll Take Away

01

Committed isn't deployed

The "$7tn" is a stack of pledges and AUM — not a deployment account.

02

Leverage scales with size

~1.8x system-wide, but 7.6x on $1bn+ deals. Small tickets are the problem.

03

Three regimes disagree

OECD, MDB Joint Report and Convergence aren't additive. Read the gap, not the sum.

01 — Anatomy of the GapCommitted vs. Deployed

The Numbers

01.1Committed Is Not Deployed

The headline "$7 trillion" is a stack of pledges, capital bases and assets-under-management — not a deployment account. The gap between what is announced and what moves is where the paradox lives.

Table 1.1 — Capital pools: announced/available vs. annual flow to EMDE development & climate
Capital pool / channel	Announced / base	Annual EMDE flow	Nature of the gap
MDB private capital mobilization	—	~$70bn	Mobilization ratio ~0.6–1.0x of own account
Blended finance market (Convergence)	$231bn cum.	~$18bn	Thin annual throughput vs. cumulative claims
GFANZ net-zero alliance AUM	~$130tn	marginal	Pledged AUM ≠ allocation to EMDE transition
Global pension & insurance AUM	~$80tn+	<1%	Fiduciary/regulatory exclusion of the asset form
Climate funds (GCF, CIF, GEF)	$10s bn	slow	Disbursement lag; accreditation bottlenecks
Philanthropy / catalytic capital	small	scarce	The binding scarce input — under-supplied
Indicative need (SDG + climate)	—	~$4tn+	Order-of-magnitude shortfall

Figures span differing base years (2023–2025) and methodologies; rows are not additive. "Announced/base" mixes capital bases, AUM and cumulative deal volume. See §Methodology for the three measurement regimes and double-counting risks.

▸The three measurement regimes

OECD

Amounts mobilized from the private sector by official interventions. Conservative, attribution-based, the basis for most cross-donor comparison.

MDB Joint Report

Splits PDM (active arranging) from PIM (anchor/parallel). Attributes PDM at full value to the arranger — flattering and double-count-prone.

Convergence

Tracks the blended finance transaction market specifically — deals using concessional capital to mobilize commercial. Hence the smaller ~$18bn figure.

Read This Carefully

Reported mobilization is best read as a lower bound on activity but an upper bound on genuinely additional private capital. Publish What You Fund's What Works (Oct 2024) argues the MDB methodology simultaneously understates emerging forms — portfolio risk transfers, balance-sheet operations — and lacks the project-level disclosure needed to verify additionality.

DENOMINATOR CHECK — The "1–2% answer" is not an artifact of picking the smallest figure. Narrow (Convergence blended-transaction market): $18bn ÷ ~$4tn ≈ 0.5%. Broad (all MDB/DFI private mobilization, 2023 Joint Report): ~$88bn ÷ ~$4tn ≈ 2.2%. Whichever regime you accept, deployed flow is a low-single-digit % of need — the paradox is robust to the denominator, not created by it.

Gravitywell Research · Confidential07 / 47

01 — Anatomy of the GapThe Frozen Ratio

The Efficiency Metric

01.2Leverage Isn't Frozen — It's Fragmented

The common claim that blended leverage is "stuck" is half-right. The average is weak — but that average hides a sharp truth: leverage rises steeply with deal size. Small bespoke tickets drag the mean down; large, aggregated, MDB-anchored structures already mobilize at 6–8x.

Average private-capital leverage ratio by transaction sizeFIG. 1.1

Private dollars mobilized per concessional dollar. Larger transactions mobilize disproportionately more — the case for aggregation and securitization.

SOURCE: Convergence — leverage by transaction size; *$500m+ climate deals with MDB/DFI presence mobilize ~5.9x (Convergence 2024). Smaller-band figures Gravitywell-interpolated from Convergence size-leverage relationship.1,2

~1.8–4x

Market-average leverage (drags on small deals)

5.9x

$500m+ climate deals, MDB present

7.6x

$1bn+ transactions

This reframes the entire problem. If leverage were structurally capped, the case for blended finance would be weak. But it isn't capped — it is fragmented across thousands of sub-scale, bespoke transactions whose near-fixed structuring costs and idiosyncratic terms crush their mobilization efficiency. The median 2024 deal was just $65m, even as a handful of $1bn+ transactions pulled the average size up.3

The strategic implication runs straight to Section 05: if large, standardized, aggregated structures mobilize at 6–8x while small ones manage barely 1x, then the binding fix is aggregation and securitization — pooling many small exposures into institutional-scale, rated instruments. The leverage potential already exists; it is trapped at the wrong transaction size. Encouragingly, concessional capital is already shifting from grants (41% of disbursement in 2022) to risk-sharing guarantees and senior positions (grants just 10% by 2024) — exactly the instruments that scale.3

SELECTION CAVEAT — Part of the size–leverage link is causal (fixed structuring costs amortize over a larger base) but part is selection: $1bn+ deals are disproportionately middle-income, hard-currency-adjacent and the most bankable — i.e. the deals least likely to need subsidy at all (§06.1). The 7.6x is therefore an upper bound on what aggregation alone can deliver for the smaller, more-additional projects that actually drive the gap.

Gravitywell Research · Confidential08 / 47

01 — Anatomy of the GapRegional Distribution

Where the Money Does — and Doesn't — Go

01.3The Geography of Under-Deployment

Capital concentrates where it is least additional. The regions with the largest gaps — Sub-Saharan Africa above all — receive a disproportionately small share of mobilized private finance, while middle-income markets absorb the bulk.

Indicative share of blended-finance flows by regionFIG. 1.2

Directional shares; SSA leads by deal count but trails on ticket size and private leverage.

SOURCE: Convergence State of Blended Finance 2024 (directional).

The concentration problem

Sub-Saharan Africa leads on the number of blended transactions but the value of mobilized private capital skews to larger, more familiar middle-income markets and to a narrow band of sectors. The result is a self-reinforcing loop: thin track records keep risk perceptions high, which keep tickets small, which keep track records thin.

Mission 300

The World Bank–AfDB initiative to connect 300 million Africans to electricity by 2030 is the flagship test of whether concessional capacity can crowd in private energy investment at continental scale. Self-reported target

▸Sectoral concentration mirrors the geography

Table 1.2 — Blended finance by sector — flow concentration & the deployment lag
Sector	Flow share	Deployment characteristic
Energy & climate	largest	Renewables IPPs bankable; grids & storage lag
Financial services / inclusion	large	MFI debt funds; securitizable, replicable
Agriculture & food	moderate	Smallholder risk; needs deep TA + first-loss
Infrastructure (non-energy)	moderate	Long tenors; FX & offtake risk; PPP-dependent
Health & education	small	Weak commercial cash flows; donor-anchored

Shares directional. The pattern: capital flows to sectors with contracted, hard-currency-adjacent cash flows (power purchase agreements) and away from social sectors and local-currency revenue models.

Gravitywell Research · Confidential09 / 47

Part I · The Deployment Gap

Section 02

The Barriers

Six interlocking constraints turn committed capital into stranded capital. The binding one is rarely the one cited — and almost never a shortage of money.

Reading

Pages 10 — 13

Binding Constraint

Bankable pipeline

What You'll Take Away

01

Regulation rejects the form

Basel, Solvency II and fiduciary rules screen out unrated, illiquid EMDE assets.

02

Pipeline is the scarcity

Bankable, investment-ready projects — not allocator appetite — is the binding limit.

03

Macro compounds it

FX mismatch, higher-for-longer USD and China's retrenchment tighten the constraint.

02 — The BarriersThe Constraint Chain

How Capital Gets Stranded

02.1Six Constraints, One Broken Pipe

Capital must clear every link in the chain to deploy. A break at any point strands the whole flow — and the system has breaks at several.

The mobilization chain — and where it breaksFIG. 2.1

SOURCE: Gravitywell Research synthesis. The pipeline link is the binding constraint; risk-pricing and fiduciary fit are secondary breaks.

① Pipeline — the binding constraint

Project preparation is chronically underfunded. Early-stage development risk — feasibility, permitting, land, offtake — is precisely the risk commercial capital will not take, and the stage where concessional money is scarcest. No pipeline, no deployment, regardless of downstream appetite.

② Regulation — Basel, Solvency II, fiduciary

Unrated, illiquid, bespoke EMDE exposures attract punitive capital charges for banks and insurers and fail the liquidity and rating screens fiduciaries apply. The form is rejected even when the credit is sound.

③ Currency risk

Project revenues are local-currency; institutional capital is hard-currency. Hedging is expensive or unavailable at tenor. TCX and local-currency facilities help at the margin; the structural mismatch persists.

④ ESG & taxonomic fragmentation

Competing taxonomies (EU, ISSB, national) and inconsistent impact metrics raise diligence cost and create greenwashing and litigation risk, deterring allocators who need defensible, comparable labels.

⑤ Transaction costs & non-standardization

Every blended deal is bespoke. Negotiation, legal and structuring costs are near-fixed, so small EMDE tickets carry ruinous unit costs. The absence of standard templates is itself a first-order barrier.

⑥ Political & ODA retrenchment

The dismantling of USAID and the US withdrawal from the FfD4 process removed a major catalytic and TA funder and chilled the policy momentum behind "billions to trillions." 2025 development

Gravitywell Research · Confidential10 / 47

02 — The BarriersRegulation & Risk Perception

The Form Problem

02.2Regulation Rejects the Wrapper, Not the Credit

The most important and least-understood barrier: EMDE development assets are often better credits than they are priced to be — but they arrive in a form that capital rules and fiduciary duty are built to reject.

▸The GEMs evidence: perception vs. reality

Realized risk vs. perceived riskFIG. 2.2

GEMs Consortium statistics on MDB/DFI private-sector lending experience.

SOURCE: GEMs Consortium (Oct 2025 release; 1994–2024 data). Default rates are count-based, through-the-cycle averages; recoveries nominal/undiscounted — not annualized hazard rates.

For decades, the default and recovery experience of MDB and DFI private-sector lending sat locked inside the Global Emerging Markets Risk Database (GEMs). When headline statistics were finally released, they showed observed private-sector default near 3.54% and recovery around 72.9% (Oct 2025 release) — risk materially lower than the perception embedded in EM pricing and credit ratings.

The Caveat That Matters

These statistics reflect the specific experience of MDBs/DFIs — institutions with preferred-creditor treatment and structuring advantages. They may not generalize to all private EMDE lending. The re-rating case is strong but not unconditional.

"The risk premium on emerging-market development debt is, in large part, an information premium. Disclose the data and a meaningful share of it should compress." Gravitywell synthesis of GEMs disclosure debate

▸How the rules bite

Table 2.1 — Regulatory & fiduciary screens that reject EMDE asset form
Regime	Mechanism	Effect on EMDE deployment
Basel III/IV (banks)	Risk-weighted capital; unrated exposures penalized	Raises cost of holding EMDE/infra loans
Solvency II (insurers)	Capital charges by asset class & rating	Long-tenor unrated infra is capital-expensive
Fiduciary duty (pensions)	Liquidity, rating, benchmark constraints	Bespoke illiquid assets fail mandate screens
Credit ratings	Sovereign ceiling caps sub-sovereign credits	Sound projects capped below investment grade

The through-line: a rated, tranched, liquid wrapper — the subject of Section 4 — can make the identical underlying credit eligible for capital that the raw loan cannot reach.

Gravitywell Research · Confidential11 / 47

02 — The BarriersPrioritization Matrix

Which Barriers To Fix First

02.3Severity vs. Tractability

Not all barriers are equal. The most valuable interventions sit where high severity meets high tractability — the upper-right quadrant — and that is where information and standardization land.

Barrier severity vs. tractabilityFIG. 2.3

SOURCE: Gravitywell Research assessment. Bubble size ∝ estimated capital unlocked if resolved.

The Cruelest Quadrant

High severity, low tractability

Bankable pipeline is the single most severe constraint — and among the hardest to fix, because project preparation is slow, local, and unglamorous. It cannot be securitized away. It requires patient, on-the-ground development capital that almost no one is paid to provide.

The Cheap Win

High severity, high tractability

Information (full GEMs disclosure) and standardization are severe barriers that are cheap to fix. They require coordination, not capital. This is why we rank them as the highest-leverage public action in the entire system.

Gravitywell Research · Confidential12 / 47

02 — The BarriersCurrency, Macro & the China Question

The Barriers the First Edition Underweighted

02.4Currency, Macro & China

Three forces shape deployment more than any structuring choice: the currency mismatch that quietly kills local-revenue projects, the post-2022 rate environment, and the largest development financier of the century — China — whose retrenchment reshaped the entire field.

▸Currency: the structural killer

The defining mismatch of development finance: institutional capital is denominated in hard currency and expects hard-currency returns, while most EMDE project revenue — power tariffs, water charges, microloans, food sales — is earned in local currency. Someone must bear the gap between the two.

Hedging exists but is expensive and short-tenor; instruments like TCX extend reach at the margin but cannot cover 20-year infrastructure exposures at acceptable cost. When hedging is unavailable, the mismatch lands on the borrower (see §06.2) or simply prevents the deal. A currency depreciation can convert a performing local-currency project into a hard-currency default with no change in the underlying business — the mechanism behind several of the failures in §04.4.

Why It Compounds Post-2022

A higher-for-longer US dollar and elevated global rates do three things at once: raise the hard-currency coupon EMDE borrowers must service, deepen local-currency depreciation pressure, and widen the spread investors demand. Many low-income countries entered this period already in or near debt distress — shrinking the fiscal space for the guarantees and first-loss that blended structures depend on. The macro backdrop is not neutral; it has been actively tightening the constraint.

▸The China question

Any global picture of development capital that omits China is incomplete. Over 2013–2021 China was the single largest bilateral development financier, deploying hundreds of billions through Belt and Road infrastructure lending — largely not blended, but direct, collateralized, state-to-state debt. That model has sharply retrenched since 2021 as China pivoted from new lending to managing a portfolio of distressed and restructured exposures across low-income borrowers.

This matters for blended finance in three ways. First, China's retrenchment removed a major source of infrastructure capital, widening the gap Western-led blended structures are now asked to fill. Second, Chinese lending created much of the existing EMDE debt overhang that constrains new borrowing today. Third, China's own institutions — the AIIB and the New Development Bank — are increasingly active in the multilateral space and are potential participants in, or competitors to, the originate-to-distribute architecture this report describes. A Western-only reading of the re-plumbing misses half the board.

Hard FX

Investor capital & return expectation

Local FX

Project revenue base

2021→

China BRI lending sharply retrenches

AIIB/NDB

New multilateral entrants & competitors

SCOPE NOTE — China's development finance is predominantly direct bilateral lending, outside the blended-finance datasets (Convergence, OECD DAC) this report relies on. Its influence on the deployment gap is therefore real but largely exogenous to the mobilization figures cited elsewhere — a structural backdrop, not a line item.

Gravitywell Research · Confidential13 / 47

Part II · What Works & What It Costs

Section 03

Instrument
Effectiveness

Eight tools. A wide range of outcomes. We score each on demonstrated leverage, replicability and where it has actually catalyzed private capital at scale — versus where it has disappointed.

Reading

Pages 14 — 17

Top Performer

Guarantees / first-loss

What You'll Take Away

01

Enhancement layer wins

Guarantees and first-loss score highest on the Effectiveness Index.

02

Highest leverage is capital-light

The best instruments mobilize the most per dollar at risk — and are undersupplied.

03

Cost-per-dollar varies widely

Mobilization efficiency differs sharply across the eight instruments.

03 — Instrument EffectivenessThe Scorecard

What Works, What Doesn't

03.1The Instrument Effectiveness Index

Each blended-finance instrument graded on demonstrated leverage, replicability and scale of catalyzed private capital — the Gravitywell Instrument Effectiveness Index. The dispersion is wide, and the highest performers are the most capital-light.

Table 3.1 — Gravitywell Instrument Effectiveness Index (leverage-weighted)
Instrument	Grade	Leverage	Where it succeeds	Where it fails / limits
Guarantees	A+	●●●●●	Capital-light; covers tail risk banks won't; very high $ mobilized per $ at risk	Chronically undersupplied; counted as exposure not deployment by issuers
First-loss / catalytic	A+	●●●●●	Junior tranche unlocks senior commercial capital; SDG Loan Fund 9:1	Scarcest input; almost entirely philanthropic/DFI; not scalable on its own
Political risk insurance	A−	●●●●○	MIGA cover decisive for frontier bankability; mobilizes cross-border equity	Coverage gaps; pricing; did not replicate at scale across markets
Technical assistance	A−	●●●●○	Builds the missing pipeline; enables everything downstream	Grant-dependent; ODA-cut-exposed; hard to attribute mobilization
Concessional debt	B	●●●○○	Improves blended cost of capital; anchors senior structures	Risk of subsidizing what would fund anyway; additionality contested
Green bonds	B	●●●○○	Scales in deep capital markets; standardized; liquid	Concentrates in middle-income issuers; weak EMDE/frontier reach
PPPs	B−	●●●○○	Aligns public/private on large infra; risk-sharing	High prep cost; contingent-liability & renegotiation risk; slow
Sustainability-linked	C	●●○○○	Flexible use of proceeds; KPI alignment	Greenwashing scrutiny; weak penalties; credibility questions

Grades and leverage ratings are Gravitywell's qualitative synthesis of demonstrated private-capital mobilization per unit of public/concessional input — editorial judgement, not a single quantitative metric. Ranked high-to-low by grade. A = high leverage/replicability · B = moderate · C = constrained.

The Paradox Within the Paradox

The two highest-leverage instruments — guarantees and first-loss capital — are the two most undersupplied. Why? Issuers count guarantees as exposure, not as deployment, so internal incentives discourage them. And first-loss capital must come from the scarcest pool of all: catalytic philanthropy and DFI equity willing to absorb the first dollar of loss.

Capital-Markets Reframe

Read as capital-markets primitives, these instruments are simply credit enhancement (guarantees, PRI), subordination (first-loss), and standardized public paper (green bonds). Section 4 shows how combining them inside securitization structures is what finally makes the package institution-grade.

Gravitywell Research · Confidential14 / 47

03 — Instrument EffectivenessWorked Example

Anatomy of a Success

03.2The SDG Loan Fund — 9:1, Decoded

The Allianz–FMO–MacArthur–Skoll SDG Loan Fund is the canonical proof that a small slice of catalytic capital, correctly placed, can mobilize commercial money at scale. Here is how the structure does the work.

SDG Loan Fund — the catalytic stackFIG. 3.1

~$1.1bn mobilized. Roughly $111m of catalytic protection unlocked ~$1bn of institutional senior capital.

SOURCE: FMO / Allianz Global Investors / MacArthur Foundation / Skoll Foundation transaction disclosures (2023–24). Figures approximate.

Why it worked

Subordination did the heavy lifting. A thin junior layer plus an unfunded guarantee made the senior tranche investment-grade-like for Allianz.
FMO originated and managed — investors bought a diversified, professionally managed EMDE loan portfolio, not single-project risk.
Catalytic capital was unfunded where possible — MacArthur's guarantee commits capital only if losses occur, maximizing leverage per dollar.

Why it is hard to replicate

The catalytic layer is the bottleneck. Scaling to $100bn needs ~$10bn of first-loss — far beyond current philanthropic supply.
Bespoke structuring took years and rare expertise; standardization (§4) is the missing ingredient for repeat issuance.

The Lesson

The instrument is proven. The constraint is the supply of catalytic first-loss capital and the absence of a standard template to repeat it cheaply.

Gravitywell Research · Confidential15 / 47

03 — Instrument EffectivenessWhere Mechanisms Disappoint

The Honest Ledger

03.3Where the Mechanisms Disappoint

A credible assessment names the failures. Three instruments routinely promoted as solutions have under-delivered against the scale claimed for them.

Underwhelms

Green bonds in frontier markets

Green bonds scale beautifully where capital markets are already deep — and barely reach the frontier markets with the largest gaps. The instrument concentrates issuance in middle-income sovereigns and corporates, doing little for the lowest-income countries.

Credibility risk

Sustainability-linked debt

KPI-linked structures with weak step-up penalties and soft targets invited greenwashing scrutiny. Issuance cooled as investors questioned whether coupons were truly at risk. Alignment without teeth is marketing.

Slow & contingent

PPPs at scale

Public-private partnerships carry heavy preparation costs, multi-year timelines, contingent-liability exposure for governments, and a long history of costly renegotiation. They work for marquee assets — rarely at portfolio scale.

▸The additionality question hangs over everything

The deepest critique cuts across all instruments: additionality. Concessional capital is only justified if it funds what commercial capital would not. Yet measurement regimes rarely verify this at the project level, and the incentive to report large mobilization numbers pushes toward claiming credit for deals that might have happened anyway.

This is why Publish What You Fund's project-level disclosure agenda matters as much as any new instrument. Without it, the sector cannot distinguish genuine catalysis from subsidized lending — and allocators cannot trust the leverage ratios that justify the public expense. The capital-markets structures in Section 4 partly answer this by making the catalytic layer explicit and priced: when the first-loss tranche is a visible, sized instrument, additionality is structural rather than asserted.

~$9.7bn

IFFIm vaccine bonds issued — a green/social-bond success

40:1

MacArthur $25m guarantee : ~$1bn mobilized (the 9:1 counts all FMO junior)

$215bn

Est. annual clean-energy equity gap, EMDEs

Synthesis

No single instrument is the answer. The pattern across successes — SDG Loan Fund, Scaling Solar, IFFIm — is combination: credit enhancement plus subordination plus standardization plus a capital-markets wrapper. The next section is about industrializing that combination.

Gravitywell Research · Confidential16 / 47

03 — Instrument EffectivenessCost Per Mobilized Dollar

For the Finance Ministry

03.4Where Scarce Public Money Works Hardest

A scorecard ranks instruments by quality; a finance ministry needs them ranked by cost-effectiveness — private capital mobilized per unit of scarce public resource committed or put at risk. On that axis, the ranking inverts the intuition that grants "do the most."

Indicative private capital mobilized per $1 of public resourceFIG. 3.2

Higher is more cost-effective. Guarantees lead because they commit capital only contingently; grants trail because the dollar is spent outright.

SOURCE: Gravitywell synthesis of Convergence size-leverage data & OECD 2025 structured-fund evidence; directional, not audited per-instrument ratios. Guarantee figure reflects contingent (unfunded) nature.1,16

Table 3.2 — Cost-effectiveness & the fiscal reality behind it
Instrument	Mobilized / public $	Why	Fiscal catch
Guarantees	Highest	Contingent — capital committed only if called	Creates contingent liabilities (see §08)
First-loss / catalytic	Very high	Thin junior layer unlocks large senior	Scarcest input; mostly philanthropic
Political risk insurance	High	Covers specific tail; capital-light	Premium pricing; coverage gaps
Concessional debt	Moderate	Improves blended cost of capital	Capital spent; additionality contested
Grants (direct)	Lowest	No leverage by design	Dollar spent outright — but builds pipeline

Grants rank last on leverage yet remain essential: they fund the project preparation and technical assistance that create the pipeline everything else depends on. Cost-effectiveness is necessary, not sufficient — sequencing matters (§05).

The Policy Instruction

For a finance ministry with a fixed concessional envelope, the cost-effectiveness ranking argues for shifting the marginal dollar toward guarantees and first-loss — which the data confirm is already happening (grants fell from 41% to 10% of disbursement, 2022–24).3 But the contingent-liability accounting in §08 must be honored, or the apparent efficiency simply hides risk off-budget.

Gravitywell Research · Confidential17 / 47

Part II · What Works & What It Costs

Section 04 · The Evidence Base

Returns, Losses
& the Re-Rating

What an investment committee actually needs: realized returns by tranche and region, the granular default and recovery record, and — kept honestly — the deals that failed.

Reading

Pages 18 — 21

Key Datum

3.54% default

What You'll Take Away

01

Fixed-income, not venture

Diversified EM/blended senior clusters at ~6–9% — not equity-like returns.

02

EM risk is over-priced

GEMs: 3.54% private default, 72.9% recovery — better than perception.

03

Disclosure could re-rate

Full GEMs data is a $600–800bn contingent "data dividend".

04 — Returns, Losses & the Re-RatingThe Return Profile

What You Actually Earn

04.1The Return Profile, Honestly

The single most important calibration for any allocator: blended-finance and EM-impact strategies are a fixed-income-like asset class with a development overlay — not a venture one. Expecting equity-like returns is the fastest route to disappointment and capital flight.

Indicative net return ranges by strategy & trancheFIG. 4.1

Net-to-investor IRR / yield bands. Tranche position drives return: senior buyers trade yield for safety and capital-charge efficiency.

SOURCE: Cambridge Associates/GIIN Impact Investing Benchmark (pooled IRR 6.9%; EM 6.2%, Africa 9.7% — 2015 vintage, dated); GIIN State of the Market 2025 (90% met/exceeded return expectations; realized PE ~16% DM / ~18% EM per 2024 ed.); J.P. Morgan CEMBI; Gravitywell tranche estimates.9,10,12,14

DATA NOTE — Self-reported impact returns run materially above the only audited pooled benchmark (6.9% — Cambridge/GIIN, but 2015-vintage, pre-2011 funds). GIIN's State of the Market 2025 reports 90% of investors met or exceeded financial-return expectations (realized PE ~16% DM / ~18% EM per the 2024 edition). The gap reflects both vintage and survivorship/selection bias. We treat the dated pooled figure as a conservative floor and survey returns as an optimistic ceiling; the true realized number sits between.9,12

The dispersion matters more than the average. The Cambridge/GIIN benchmark's headline 6.9% pooled IRR sat just below the 8.1% of a non-impact comparison set — but emerging-market and Africa-focused subsets performed in line or better (Africa funds returned 9.7%). The lesson for allocators is that selection dominates: the asset class can deliver competitive risk-adjusted returns, but manager and structure choice drive the outcome far more than the "blended" label.

For the senior tranches this report steers most allocators toward, the relevant comparison is not equity but rated EM credit. A senior note yielding 3–6%, capital-charge-efficient under Solvency II and backed by a diversified, GEMs-evidenced pool, competes against investment-grade EM corporates — not against private equity. Underwritten on that basis, the proposition is coherent. Underwritten as a return-maximizing play, it is not.

Gravitywell Research · Confidential18 / 47

04 — Returns, Losses & the Re-RatingThe Loss Record

The Default & Recovery Evidence

04.2The Loss Record, Decomposed

The 2024–25 GEMs releases are the most important dataset in development finance. Drawn from 1994–2024 lending by MDBs/DFIs to 10,000+ private entities across 169 countries, they let risk be priced from evidence rather than perception.4,6

Table 4.1 — GEMs default & recovery, by counterpart and segment (1994–2024)
Segment	Avg. default rate	Avg. recovery	Reference comparison
Private entities (all)	3.54%	72.9%	≈ advanced-economy sub-IG firms; median recovery >90%
— Financial sector	2.26%	79.1%	Lowest default, highest recovery — most securitizable
— Sub-Saharan Africa	6.05%	>78%	≈ S&P "B" to "B−"; highest recovery of any region
— Low-income countries	6.3%	—	Gap to advanced economies smaller than perceived
State-owned enterprises	2.6%	88%	Publicly-guaranteed; low volatility
Sovereign & sov.-guaranteed	0.77–1.06%	94.9%	Far below CRA sovereign default observations

Figures from GEMs Consortium publications (Oct 2024 & Oct 2025). Private all-segment default reported at 3.54%; recovery 72.9% (a 2024 release reported 72.2%). Median private recovery exceeds 90% — the mean is dragged by a tail of low-recovery cases. Caveat: GEMs default rates are count-based, through-the-cycle averages and recoveries are nominal/undiscounted — not annualized hazard rates. The "≈ S&P B/B−" mapping holds only on the cohort-adjusted marginal basis IFC constructed (2024 note), and even then the comparator pool is more advanced-economy-weighted than GEMs.4,5,6,7

The perception gapFIG. 4.2

SSA private default vs. how its sovereign ratings imply it should price.

SOURCE: GEMs (SSA private 6.05%); perceived band illustrative of sovereign-ceiling pricing.4,8

Three things the data settle

EMDE private lending is not exotic-risk. A 3.54% blended default rate sits alongside advanced-economy non-investment-grade corporates — a familiar, priceable band.
Recovery is the hidden strength. Mean recovery near 73% — and a median above 90% — beats many DM benchmarks, reflecting MDB workout capacity and seniority.
The sovereign ceiling is the distortion. Sub-sovereign and private projects are routinely capped by their sovereign's rating, even though GEMs shows their default experience is materially better.

The Honest Caveat

These are MDB/DFI outcomes — preferred-creditor treatment, intensive structuring and workout. A private lender without those protections should haircut these figures, not adopt them wholesale.

Gravitywell Research · Confidential19 / 47

04 — Returns, Losses & the Re-RatingThe $600–800bn Dividend, Derived

From Data to Capital

04.3The Data Dividend, Derived

In the first edition we cited a "$600–800bn data dividend." Here we attribute and decompose it. The figure originates with the Center for Global Development; the mechanism is a re-rating of capital charges and required risk premia once realized loss data is public and rating agencies act on it.6

The transmission: from disclosure to deployed capitalFIG. 4.3

SOURCE: Center for Global Development (dividend estimate); G20 MDB Roadmap progress & MDB Reform Accelerator (CRA posture, GEMS 2.0).6,21,22

How CGD Frames It

The estimate is an order-of-magnitude sizing of the capital that improved transparency could mobilize as ratings and capital rules move toward realized rather than perceived EMDE risk. It is a directional re-pricing opportunity, not an audited transaction total — and it is contingent on rating agencies and regulators actually responding.6

Why 2025 Is the Inflection

Two facts move this from theory toward practice: GEMS 2.0 will spin out as a standalone entity with full MDB backing, and the CAF process reports that at least one credit rating agency has signaled willingness to revisit its methodology in light of MDB credit-performance data. The transmission channel is opening.21,22

Table 4.2 — Data-dividend sensitivity: the unlock is contingent on who responds (Gravitywell scenarios)
Scenario	What actually happens	Near-term capital unlocked
Low	GEMs disclosure improves analytics; CRA/Basel/Solvency capital rules unchanged	~$50–100bn
Base	CRAs grant partial callable-capital / portfolio credit; regulators move slowly	~$200–350bn
High	Full CRA re-rating and capital-charge response (the CGD / S&P upper bound)	$600–800bn
Prob-weighted	Our central case — directionally the largest single lever in the system	~$150–250bn

The $600–800bn (CGD Note 385, Dec 2025; S&P MDB-headroom model) is the High case, not a forecast. The dividend materializes only to the extent rating agencies and prudential regulators act on the data; absent that, disclosure improves analytics without moving capital charges. Scenario splits are Gravitywell estimates.6

Gravitywell Research · Confidential20 / 47

04 — Returns, Losses & the Re-RatingWhen It Goes Wrong

The Other Side of the Ledger

04.4What Failure Looks Like

A registry of only winners cannot be trusted. The re-rating thesis depends on losses being real but contained — so the honest test is what happens when deals don't work. Here is where blended finance underperforms, fails, or destroys value.

Structural failure

The unrated-tranche trap

In many structured funds the senior and mezzanine tranches are unrated. If credit losses deplete the first-loss equity — or the underlying worsens — there is no rating to re-rate, and the senior holder cannot mark, hedge or exit cleanly. Amundi's 2025 analysis flags this as a core reason credit enhancement fails to mobilize at scale: the wrapper is incomplete without a rating.11

Sector failure

Off-grid energy write-downs

A wave of pay-as-you-go solar and mini-grid ventures across East Africa absorbed concessional and impact equity in the late 2010s; several restructured, were written down, or exited below cost as currency depreciation and collection rates undercut local-currency revenue models. The pattern foreshadowed the FX lesson of §02.13

Underperformance

The impact-return drag

The only audited pooled benchmark (Cambridge/GIIN, 2015-vintage) underperformed its non-impact comparator (6.9% vs. 8.1% IRR). It is dated, and current self-reported returns run higher — but the headline holds: across the full set, the impact overlay coincided with a return give-up. Allocators who under-diligence the manager inherit the drag, not the outperforming tail.9,10

Renegotiation risk

PPP contingent liabilities

A material share of emerging-market PPP infrastructure contracts are renegotiated post-award, frequently shifting cost and risk back to the public balance sheet and triggering contingent liabilities that were never budgeted. The instrument's headline "risk-sharing" can invert under stress — see the fiscal accounting in §08.

Where value leaks — failure modes mapped to the capital stackFIG. 4.4

SOURCE: Gravitywell Research synthesis of Amundi (2025) credit-enhancement analysis and documented sector write-downs.11,13

The Takeaway That Protects Capital

Failures cluster at three points: missing ratings (no exit), local-currency revenue (FX wipes margin), and under-sized first-loss (senior absorbs losses it was promised it wouldn't). Every one is diligenceable before commitment. The asset class is not uniquely dangerous — but it is uniquely punishing of lazy structuring.

Gravitywell Research · Confidential21 / 47

Part III · The Re-Plumbing

Section 05 · The Differentiated Angle

The Capital-Markets
Re-Plumbing

Securitization. Originate-to-distribute. Synthetic risk transfer. The GEMs data dividend. This is where billions become trillions — by replacing the broken pipe with plumbing the buy-side already knows how to use.

Reading

Pages 22 — 25

Flagship

IFC MCPP

What You'll Take Away

01

Originate-to-distribute

MDBs recycle balance sheets by selling or securitizing loans.

02

Securitization institutionalizes

Bespoke loans become rated, tranched, tradable senior paper.

03

Standardize per sector

Contracts + loss data are what make a sector securitizable.

05 — The Capital-Markets Re-PlumbingOriginate-to-Distribute

The New Operating Model

05.1From Balance-Sheet Lending to Originate-to-Distribute

The reform that matters most is not a new fund — it is a new business model. MDBs and DFIs are learning to originate loans, retain a sliver of risk, and distribute the rest to institutional balance sheets. Capital recycles; leverage compounds.

Balance-sheet model vs. originate-to-distributeFIG. 5.1

SOURCE: Gravitywell Research. Leverage figures illustrative of the model's potential, not a single program's realized result.

In the legacy model, an MDB lends a dollar, holds it to maturity, and that dollar is unavailable to lend again for years. Leverage on third-party capital is roughly one-to-one or less. The balance sheet is the binding constraint on volume.

In originate-to-distribute, the MDB uses its origination franchise and structuring expertise to create assets, retains only the junior or first-loss slice that its concessional mandate is suited to hold, and sells the senior tranche — rated and de-risked — to insurers, pensions and asset managers. The freed capital is recycled into new origination. Balance-sheet velocity, not balance-sheet size, becomes the driver. This is the single most important behavioural change the CAF reforms (§6) are trying to institutionalize.

Why The Buy-Side Can Finally Participate

A rated, tranched senior note backed by a diversified pool of MDB-originated loans clears the Basel, Solvency II and fiduciary screens that the raw, unrated, single-name loan never could. The credit is the same. The wrapper is what changed — and the wrapper is the whole game.

VELOCITY, WORKED — Hold-to-maturity: $100 of capital makes $100 of loans, once. Under O2D the MDB lends $100, distributes the ~$85 senior, retains the ~$15 junior, recycles the $85 — but each new cycle still consumes ~$15 of junior, so the binding constraint is junior-capital capacity, not origination. With a fixed $100 of loss-absorbing capital at a 15% retention, the model supports ≈$670 of cumulative origination (~6.7×) before the junior is exhausted. Velocity is real, but it is capped by the first-loss slice — which is why §08's "catalytic-capital supply" is the swing variable, not balance-sheet size.

Gravitywell Research · Confidential22 / 47

05 — The Capital-Markets Re-PlumbingSecuritization Mechanics

How The Structures Actually Work

05.2Tranching, True-Sale and Synthetic Risk Transfer

Three structures dominate the re-plumbing. Conflating them overstates maturity. Here is the precise mechanics — and the critical distinction between moving the asset and moving the risk.

The capital structure waterfallFIG. 5.2

SOURCE: Gravitywell Research. Indicative tranche sizing.

True-sale (cash) securitization

Assets are legally sold into a special-purpose vehicle, which issues rated notes. Capital genuinely leaves the originator's balance sheet. IFC's 2025 ~$510m CLO is the landmark true-sale cash structure — the first of its kind for the institution.

Synthetic risk transfer (SRT)

The originator keeps the assets but transfers credit risk via guarantees, insurance or credit-default swaps. Regulatory capital is freed without selling the loans. Room2Run (AfDB / Mariner) and IDB Invest's Scaling4Impact ($1bn, now >$1.3bn) are synthetic.

Critical Distinction

Room2Run and Scaling4Impact move risk; the IFC CLO moves the asset. Synthetic frees regulatory capital but keeps the assets and funding on balance sheet (it needs a protection-seller). True-sale frees both funding and capital but needs a cash buyer and a true-sale legal opinion — which is why true-sale scales to capital markets while synthetic scales to bank/insurer risk appetite.

The Co-Investment Economics

Tranching is what makes one asset serviceable to many investors with different risk mandates. The pension fund buys senior at low yield with low capital charge; the PE/infrastructure fund buys mezzanine for spread; the DFI and philanthropy hold first-loss because their mandate is catalysis, not return. Each investor buys the risk slice their rules and return targets permit — which is exactly why a structure can mobilize capital that no single-tranche instrument could.

▸Worked example — IFC's inaugural CLO (2025), and how the senior stays safe

Table 5.2 — IFC EM Securitization 2025-1 ($510m; 57 companies, 28 countries)
Tranche	Size	Attach–detach	Rating / holder
Senior (Class A)	$320m · 62.7%	37.3–100%	Aaa · sold to institutions (SOFR+130)
Mezzanine	$130m · 25.5%	11.8–37.3%	Retained by IFC
First-loss / equity	$60m · 11.8%	0–11.8%	Retained by IFC

The loss waterfall, run

Credit enhancement below the Aaa senior = 37.3% (mezz + equity). Losses hit bottom-up: the first 11.8% of pool losses are absorbed by the equity, the next 11.8–37.3% by the mezzanine — and the senior loses its first dollar only once cumulative pool losses exceed 37.3%. Against a GEMs gross-loss expectation near ~1% a year (3.54% default × ~27% loss-given-default), 37.3% is a >30× buffer. That is why a pool of sub-IG EM loans produces a genuinely Aaa senior note — the rating is bought with subordination, not optimism.

SOURCE: IFC / World Bank, EM Securitization 2025-1 (closed 19 Sep 2025; arranger Goldman Sachs; LSE-listed). Capital note: even Aaa securitization positions carry a regulatory floor (Basel SEC-SA risk-weight floor 15%, or 10% if STS-qualifying), so the wrapper is efficient but not free.24

Gravitywell Research · Confidential23 / 47

05 — The Capital-Markets Re-PlumbingStructure Registry

The Re-Plumbing, Catalogued

05.3Registry of Landmark Structures

The programs industrializing originate-to-distribute, by structure type and status. Note where figures are self-reported or forward-looking. (For the same deals seen as sector case studies, §07; as a mobilization track record, §09.6.)

Table 5.1 — Landmark capital-markets structures in development finance
Program / deal	Sponsor	Structure	Scale	Status / note
MCPP (Managed Co-Lending)	IFC	Syndication / portfolio co-lending	$16–25.5bn	Flagship; figure varies by report date
IFC CLO (inaugural)	IFC / World Bank	True-sale cash CLO	~$510m	Closed 2025; first of its kind milestone
Room2Run	AfDB / Mariner	Synthetic securitization	~$1bn ref.	Pioneer SRT; balance-sheet optimization
Scaling4Impact	IDB Invest	Synthetic securitization	$1bn→$1.3bn	LatAm/Caribbean landmark; 2024
IABS series	Bayfront (Clifford Capital)	Infra debt securitization	multi-deal	Asia infra ABS; recurring issuer
Multi-originator SST	AfDB-led	Synthetic (exploratory)	TBD	Dec 2024 LOI to explore forward-looking
MIGA Guarantee Platform	World Bank Group	Consolidated guarantees / PRI	$12.3→$20bn	By-2030 target self-reported
SDG Loan Fund	FMO / Allianz	Tranched fund + guarantee	~$1.1bn	9:1 private:public; 2023–24

Conflicting datapoints noted: MCPP cumulative totals are reported at $16bn, $19bn and $25.5bn across different dates; we present the range. The AfDB multi-originator SST was a Letter of Intent to explore a structure, not a priced deal. See §Methodology.

Cumulative landmark structure volume — the curve is bendingFIG. 5.3

Illustrative cumulative volume of MDB/DFI capital-markets structures. The acceleration post-2023 is the signal.

SOURCE: Gravitywell synthesis of sponsor disclosures. Indicative; not to scale of any single audited series.

Gravitywell Research · Confidential24 / 47

05 — The Capital-Markets Re-PlumbingFrom Pilot to Platform

The Missing Ingredient

05.4From Pilot to Platform

Every structure in §05.3 works — as a bespoke, hand-built one-off. None is yet a repeatable asset class. What separates a pilot from a platform is the unglamorous infrastructure that lets the next deal copy the last one cheaply. (The re-rating prize that disclosure unlocks is sized separately in §04.3.)

The Bottleneck

The constraint is no longer proof-of-concept — it is standardization plus catalytic-capital supply. Until documentation, ratings and loss-data are common goods, every deal pays first-of-its-kind costs; and until first-loss capital scales, the senior tranches have nothing to sit on.

What's still missing

Standardized documentation — common term sheets and legal templates so the next CLO/SRT is a copy, not a custom build. The single biggest cost-down lever.
Repeat issuance & a secondary market — Bayfront's recurring IABS series shows the model; most sponsors are still doing one-offs with no secondary liquidity.
Public loss data & ratings on tap — the disclosure that makes tranches rateable without bespoke diligence each time (§04.3).
A deeper catalytic / first-loss pool — the binding ceiling on how much senior can be built (§08.2).

Pilot → platform: the maturity chainFIG. 5.4

SOURCE: Gravitywell Research framework.

Who's Building It

The World Bank Private Sector Investment Lab (BlackRock, AXA, HSBC and other buy-side principals) embodies the shift: the demand side is now co-designing standardized, rateable supply. Bayfront's recurring IABS programme and IFC's securitization platform (§05.2) are the first true platforms rather than pilots — the template the rest of the market still has to copy.

Gravitywell Research · Confidential25 / 47

Part IV · Scrutiny, Evidence & Outlook

Section 06 · The Hard Questions

Counterfactual
& Critique

Does blended finance actually add anything — or subsidize deals that would have closed anyway? And what does the whole apparatus look like from the borrower's chair, not the allocator's?

Reading

Pages 26 — 27

The Tension

Additionality

What You'll Take Away

01

Additionality is the test

Did the support create investment that wouldn't have happened otherwise?

02

Not full crowding-out

Evidence rejects total crowd-out — but substantial additionality is thin.

03

The borrower's view

Cost and conditionality look different from the receiving end.

06 — Counterfactual & CritiqueThe Additionality Problem

The Deepest Critique

06.1Is Any of This Additional?

The most damaging question in development finance is not "does it work?" but "would it have happened anyway?" If concessional capital cheapens deals commercial money would have funded regardless, it isn't catalysis — it is a subsidy that crowds out the very private capital it claims to crowd in.

The Evidence, Stated Plainly

The most rigorous econometric review the OECD has published (Dimos & Pugh) rejects complete crowding-out of private investment by public subsidy — but finds no evidence of substantial additionality either.15 The honest reading: blended finance is neither the pure catalyst its champions claim nor the pure subsidy its critics allege. It sits in an uncomfortable, under-measured middle.

The additionality spectrum — and where the evidence places the fieldFIG. 6.1

SOURCE: Dimos & Pugh, via OECD Development Co-operation Working Paper 91 (2021).15

The field is responding. The OECD's 2025 DAC Blended Finance Guidance elevates minimum concessionality — using no more public subsidy than necessary to address a specific market failure — to a core principle, precisely to limit crowding-out.16,18 The IFC-led DFI Working Group issued enhanced additionality guidelines in 2025 requiring deal-level justification of the market failure addressed and evidence that the concessional element is the minimum needed.17 BII rates investor contribution on a four-point scale and rejects deals deemed immaterial.19

This is where the capital-markets framing earns its keep analytically. When the catalytic layer is an explicit, sized, priced first-loss tranche or guarantee — rather than a blended interest-rate subsidy buried in a loan — additionality becomes structural and visible: the public contribution is exactly the loss-absorbing capacity that the private senior would not provide. It does not prove additionality, but it makes it auditable in a way that concessional pricing never is. Mazzucato's critique — that financial engineering at the margins cannot realign private capital with public missions — is a caution, not a refutation: structures must be designed around clearly defined missions, not as ends in themselves.20 The sharper objection is Daniela Gabor's "Wall Street Consensus": that the "de-risking state" absorbs demand, currency and political risk so institutional investors keep the upside — socializing downside while privatizing reward, and entrenching a subsidy-dependent asset class. We take it seriously, and it is precisely why this report insists the catalytic layer be explicit, sized, priced and additionality-tested rather than a hidden rate subsidy — and why the GEMs evidence (that EM risk is over-priced) matters: if the de-risking is correcting a genuine mispricing, it is catalytic; if it is underwriting risk private capital should bear at market price, Gabor is right. The data says it is closer to the former, but the discipline must be enforced deal by deal. The full signed bear case — the critique at full strength, with our answer — is in §09.10.

MEASUREMENT HONESTY — Every mobilization figure in this report inherits the additionality problem: it counts capital that moved alongside public money, not capital that moved because of it. Treat all leverage ratios as upper bounds on genuine catalysis.

Gravitywell Research · Confidential26 / 47

06 — Counterfactual & CritiqueThe Borrower's-Eye View

From the Other Chair

06.2The Borrower's-Eye View

This report — like the entire industry literature — is written from the allocator's chair. That is itself part of the problem. Seen from the project developer or recipient-country institution, the same system looks slower, costlier and more extractive than the mobilization numbers suggest.

What the borrower actually experiences

Time-to-close measured in years. Bespoke blended structures require negotiating multiple institutions' legal, ESG and governance requirements in sequence. A developer's capital need is immediate; the structure's timeline is not.
Transaction costs that punish small tickets. Near-fixed legal and structuring costs make sub-$25m deals uneconomic to blend — exactly the size most local projects need.
Standardization can exclude. The templates this report champions (§05) speed up capital — but a standard PPA or microfinance template can leave genuinely novel or local-context projects with nowhere to fit.
FX risk lands on the borrower. When hard-currency capital meets local-currency revenue, the developer often bears the mismatch the allocator refused.

The Standardization Paradox

Standardization is simultaneously the report's central recommendation and a genuine risk to the borrowers it aims to serve. The resolution is tiered: standardize the high-volume, commoditizable segments (MFI debt, solar IPPs) to free capital and attention for the bespoke, high-additionality projects that actually need human structuring. Standardization should concentrate scarce structuring capacity on hard cases, not erase them.

Domestic Capital — The Missing Half

Nearly the entire blended-finance discourse concerns foreign institutional capital. For many countries the more durable answer is deepening local pension, insurance and bank capital — which carries no FX mismatch and builds permanent market infrastructure. Foreign mobilization is a bridge; domestic resource mobilization is the destination.

Table 6.1 — The same transaction, two perspectives
Dimension	Allocator's view	Borrower's view
Speed	Diligence protects capital	Years of delay; need is now
Cost	Structuring fees justified by risk	Fixed costs make small deals unviable
Standardization	Enables scale & securitization	May exclude non-template projects
Currency	Hedged or avoided	Mismatch often borne locally
Success metric	Private capital mobilized	Capital actually received, on time

A report that optimizes only the left column will keep producing impressive mobilization figures and disappointing development outcomes. Stakeholders should demand both columns be measured.

Gravitywell Research · Confidential27 / 47

Part IV · Scrutiny, Evidence & Outlook

Section 07

Transaction Registry
& Case Studies

Named deals, real structures, documented outcomes — across climate, renewable energy, infrastructure, healthcare, financial inclusion and agriculture. What replicated, and what did not. The sector cut; the structure-type registry is §05.3, the mobilization track record §09.6.

Reading

Pages 28 — 30

Sectors

Six verticals

What You'll Take Away

01

Named, sourced deals

Real transactions across energy, health, inclusion, infrastructure and agri.

02

Inclusion & energy lead

The most securitizable sectors today — granular, data-rich, standardized.

03

Equity is the bottleneck

Debt structures mature fast; a ~$215bn EMDE clean-energy equity gap remains.

07 — Transaction RegistryEnergy & Climate

Case Studies · Energy & Climate

07.1Where Renewables Became Bankable

ReplicatedEnergy

Scaling Solar

A World Bank Group "one-stop-shop": standardized tenders, template PPAs, IFC financing and MIGA political-risk cover packaged together. The standardization is the innovation — it drove record-low solar tariffs in Zambia and beyond and proved frontier solar IPPs bankable. The package, not any single instrument, did the work.

Lesson: standard documents + bundled de-risking = repeatable bankability. The template is the asset.

ReplicatedClimate

Climate Investor One

A blended facility (Climate Fund Managers / FMO) combining a development fund, construction-equity fund and refinancing fund in one vehicle — addressing the riskiest early stages where commercial capital won't go. By financing the whole project lifecycle, it tackles the pipeline constraint directly.

Lesson: fund the development stage explicitly, or there is no asset to securitize later.

Table 7.1 — Energy & climate transaction registry
Transaction	Lead	Mechanism	Outcome
Scaling Solar PRI	WBG / IFC / MIGA	Standard tender + PPA + PRI	Record-low tariffs; replicated across markets
Climate Investor One Fund	CFM / FMO	Lifecycle blended fund	Mobilized construction-stage equity
SDG Loan Fund (energy sleeve) Fund	FMO / Allianz	Tranched fund + guarantee	9:1 leverage; institutional senior
Mission 300 Concessional	WBG / AfDB	Concessional + private crowd-in	300m connections by 2030 forward

The Equity Gap Nobody Funds

Debt structures are maturing fast. Equity is the bottleneck. We estimate a ~$215bn annual clean-energy equity gap in EMDEs (Gravitywell estimate, derived from the IEA's ~$0.9–1.1tn/yr EMDE private clean-energy need, of which equity is the binding minority slice; IEA World Energy Investment 2025) — the part first-loss debt and senior notes cannot fill. This is the clearest opening for SWFs, infrastructure funds and PE described in Section 6.

Gravitywell Research · Confidential28 / 47

07 — Transaction RegistryHealth · Inclusion · Infrastructure

Case Studies · Social & Infrastructure

07.2Healthcare, Inclusion & Infrastructure

Health

IFFIm

The International Finance Facility for Immunisation converts long-dated donor pledges into immediate, tradeable "vaccine bonds" — ~$9.7bn issued. A genuine capital-markets success: it front-loads aid by securitizing the donor promise itself.

Health

Health Finance Coalition

Blended vehicles channeling catalytic and commercial capital into African health enterprises and supply chains — addressing the weak commercial cash flows that keep health donor-dependent.

Inclusion

MFI debt funds

Microfinance loan portfolios are diversified, granular and securitizable — among the most replicable blended assets. Standardized MFI debt funds are a template for what other sectors could become.

▸Financial inclusion — the securitization template

Financial inclusion is the sector where the capital-markets model is most mature. Microfinance portfolios share the properties securitization rewards: many small, statistically independent exposures with observable loss histories. That granularity makes them poolable, tranchable and rateable — the closest thing in development finance to a commodity asset class.

The lesson generalizes. Sectors become securitizable as they accumulate standardized contracts and loss data. Energy got there via standard PPAs; inclusion via standard microloans. Agriculture and health remain bespoke — which is precisely why they lag. The policy implication is to invest in standardization and data per sector, not just in capital.

Table 7.2 — Social-sector & inclusion transaction registry
Transaction	Lead / partners	Mechanism	Replicability
IFFIm vaccine bonds Securitization	Gavi / World Bank	Securitized donor pledges	High — proven, repeat issuance
MFI debt funds Fund	DFIs / impact managers	Pooled microloan portfolios	High — granular & data-rich
Health Finance Coalition Catalytic	Philanthropy / DFIs	Blended catalytic vehicles	Moderate — cash-flow-constrained
Infra debt (IABS) Securitization	Bayfront	Securitized infra loans	Growing — recurring issuer

Institutional Roles, Mapped

World Bank / IFC — origination, MCPP, CLO, Scaling Solar. Regional banks (AfDB, ADB, IDB Invest, EBRD) — SRT pioneers, regional pipelines. GCF — concessional anchor (disbursement-constrained). DFC / former USAID — guarantees & TA (now politically diminished). BII, Proparco, FMO — European bilateral originators and fund sponsors. Rockefeller, Gates, MacArthur (Catalytic Capital Consortium) — the first-loss layer.

Gravitywell Research · Confidential29 / 47

07 — Transaction RegistryAgriculture

Case Studies · Agriculture & Land Use

07.3The Hardest Sector to Bank

Agriculture concentrates every barrier at once: smallholder credit risk, local-currency revenues, climate exposure and thin data. It is the clearest test of whether blended finance can reach the genuinely difficult.

Agri

&Green Fund · the carbon cash flow

Targets deforestation-free supply chains with catalytic capital. Increasingly the cash flow that makes land-use bankable is carbon: with Paris Article 6.4 now operational (rules finalized COP29–30, 2024–25) and the ICVCM Core Carbon Principles policing integrity, contracted carbon-credit revenue — ERPAs, often with price floors — can service debt or sit as a first-loss cushion, turning a standing forest into a financeable asset.

Agri

AgDevCo

A specialist agriculture impact investor providing patient capital and hands-on development to African agribusinesses — effectively a pipeline factory, building the investable enterprises that later structures can finance.

Why agriculture lags — barrier concentration by sectorFIG. 7.1

Relative barrier intensity. Agriculture (rightmost) carries the heaviest combined load.

Credit / offtake riskFX mismatchData scarcityPreparation cost

SOURCE: Gravitywell Research assessment. Relative, illustrative.

The Sequencing Insight

You cannot securitize what has not been originated. Agriculture's path runs through pipeline-building specialists (AgDevCo) and catalytic structuring (&Green) first — accumulating standardized contracts, loss data and, increasingly, a carbon-revenue offtake (Article 6 / VCM) as the contracted cash flow that anchors repayment — before capital-markets structures can follow. The order is non-negotiable: TA and first-loss, then standardization, then securitization.

Gravitywell Research · Confidential30 / 47

Part IV · Scrutiny, Evidence & Outlook

Section 08

Outlook &
Investment Implications

The next decade, the reform agenda that would accelerate it, and a decision-ready playbook for every type of allocator — pension, insurer, SWF, PE, infrastructure fund and government.

Reading

Pages 31 — 34

Base Case

Institutionalize, not transform

What You'll Take Away

01

Institutionalize, not transform

Base case: a bigger, more professional market — still short of need.

02

Catalytic capital is the swing

First-loss supply gates every scenario from bull to bear.

03

Playbook by investor type

Distinct positioning for pensions, insurers, SWFs, PE and government.

08 — Outlook & ImplicationsThe Reform Agenda

What Would Actually Move The Needle

08.1The Unfinished Reform Agenda

The G20 Capital Adequacy Framework review identified hundreds of billions in latent MDB lending headroom. The easy reforms are underway. The hard ones — the ones that matter most — are not.

Table 8.1 — Reform status ledger
Reform	What it does	Status
CAF balance-sheet optimization	Frees lending headroom — >$650bn over the decade (MDB Heads, Jun 2025; >$250bn added since mid-2024); ~$300–400bn near-term CAF (G20 2025)	Underway progressing
Hybrid capital instruments	New capital without new shareholder cash	Early issuance progressing
GEMs full disclosure	Re-rates the asset class (§4.4)	Headlines only stalled
Standardized documentation	Cuts transaction cost; enables repeat issuance	Fragmentary stalled
Callable capital mobilization	~$1.2tn pledged but never called — huge latent backing	Unimplemented hardest
FfD4 Sevilla Commitment	Global FfD framework + Sevilla Platform for Action (130 initiatives; debt "pause clause" alliance)	Concluded Jul 2025; US withdrew weakened
SDR rechanneling	Redirects IMF reserves to development lending	Limited progress

CAF headroom estimates vary widely by source — from $60bn to $572bn (bottom-up studies) to ~$300–400bn (G20 2025 near-term CAF) to >$650bn total over the decade (MDB Heads Joint Statement, Jun 2025, >$250bn added since mid-2024). We disclose the range rather than reconcile it. The figures are projections, not realized lending.

The Easy Reforms (Happening)

Balance-sheet optimization, hybrid capital and the securitization programs of Section 4 are advancing because they require technical change within existing mandates. These grow the market.

The Hard Reforms (Stalled)

Full GEMs disclosure, true standardization and callable-capital mobilization require collective political action and the surrender of institutional control. These are what would deliver trillions — and they are stuck.

"Callable capital — ~$1.2tn pledged but never drawn — is a default backstop, not lending headroom. It cannot be deployed; the live question is how much of it rating agencies will let MDBs prudently count toward capacity. Even a modest credit would expand capacity more than any concessional pledge." Gravitywell synthesis of CAF review literature

Gravitywell Research · Confidential31 / 47

08 — Outlook & ImplicationsDecade Scenarios

Three Paths To 2035

08.2How The Market Evolves

Our base case is institutionalization, not transformation: the market grows and professionalizes, but falls short of trillions absent a step-change in catalytic-capital supply or a regulatory re-rating.

Annual private mobilization — three scenarios to 2035FIG. 8.1

SOURCE: Gravitywell Research scenarios. Illustrative trajectories, not forecasts.

Bull · Re-Rating

Full GEMs disclosure + standardization + callable-capital mobilization trigger a regulatory re-rating. Securitization scales; institutional capital floods senior tranches. Trillions become plausible.

SHIFT TO BULL IF → loan-level GEMs disclosure published and ≥1 rating agency re-rates DFI senior tranches and ≥$25bn standardized issuance by 2028.

Base · Institutionalize

Securitization programs grow steadily; the market professionalizes and roughly doubles or triples. But catalytic-capital scarcity caps leverage. A bigger market — still short of need.

HOLDS WHILE → annual blended market stays in the $18–40bn range and the system-wide leverage ratio remains below ~3x.

Bear · Retrenchment

ODA cuts deepen, political momentum dissipates, first-loss capital dries up. The re-plumbing stalls for lack of the junior capital that makes senior tranches safe. Stagnation.

SHIFT TO BEAR IF → catalytic/first-loss commitments fall YoY for two consecutive years or >3 major donors cut ODA >15%.

The Swing Variable

Every scenario turns on one input: the supply of catalytic first-loss capital. It is the keystone that makes senior tranches investment-grade. Expand it — via philanthropy, SDR rechanneling, or DFI equity — and the bull case opens. Starve it, and the most elegant securitization has no foundation to sit on.

Gravitywell Research · Confidential32 / 47

08 — Outlook & ImplicationsThe Fiscal Fine Print

For Treasuries & Debt Managers

08.3The Contingent-Liability Reckoning

The report's central policy recommendation — shift public money toward guarantees — carries a fiscal catch that the cost-effectiveness charts conceal. A guarantee is "capital-light" only until it is called. Treating contingent exposure as free is how risk migrates off-budget and surfaces later as crisis.

Why guarantees look free — and aren't

An unfunded guarantee commits no cash at signing, so it mobilizes private capital at a very high apparent ratio and often escapes headline debt and deficit metrics. But it creates a real, probability-weighted future obligation. If the guaranteed exposure defaults, the public guarantor pays — frequently at the worst moment, when multiple exposures correlate in a downturn or currency crisis.

The 2025 G20 CAF work explicitly grapples with the analogous question at MDB level: how much callable capital (~$1.2tn pledged but never drawn) can prudently count toward lending capacity, and how rating agencies should treat it. The recommendation to incorporate a prudent share of callable capital into capital-adequacy metrics remains unimplemented as of 2025 — precisely because quantifying contingent backing is hard and consequential.21,23

The Discipline Required

A credible guarantee program must: (1) price the contingent liability actuarially, not treat it as zero; (2) budget an expected-loss provision against it; (3) disclose the gross exposure and its correlation structure; and (4) stress-test for the scenario where many guarantees are called at once. Skip these, and the "most cost-effective instrument" becomes a hidden sovereign liability.

Guarantee economics — headline vs. fiscally honestFIG. 8.2

SOURCE: Gravitywell Research. Illustrative; the point is that honest accounting nets expected loss and correlated tail risk against the headline mobilization.

Reconciling §03.4 and §08.2

There is no contradiction: guarantees are the most cost-effective mobilization tool and they carry real contingent liabilities. The instruction is to use them aggressively while accounting for them honestly — provisioning, disclosing and stress-testing — so that cost-effectiveness reflects true risk-adjusted value rather than an accounting artifact.

Gravitywell Research · Confidential33 / 47

08 — Outlook & ImplicationsThe Allocator Playbook

Decision-Ready · By Investor Type

08.4The Allocator Playbook

The single most useful output of this report: where each type of institution should look, what risk slice fits its mandate, and what to demand from sponsors.

Table 8.2 — Positioning by investor type
Investor type	Where to play	Risk slice	What to demand
Pension funds	Rated senior tranches of O2D / MCPP / CLO programs	Senior	External rating; liquidity terms; GEMs-backed loss data
Insurers	Senior infra debt securitizations (IABS-type); long tenor match	Senior	Solvency II–efficient structuring; cash-flow predictability
Sovereign wealth funds	Mezzanine + anchor equity; clean-energy equity gap	Mezz / equity	Co-investment rights; pipeline visibility; governance seats
PE & infra funds	Mezzanine spread; development-stage & construction equity	Mezz / equity	DFI co-investment; PRI cover; exit/refinancing path
Asset managers	Diversified EMDE debt funds; managed portfolios (SDG-Fund model)	Senior / mezz	Professional origination; track record; standard docs
Governments / DFIs	Supply first-loss, guarantees, TA; mandate O2D & disclosure	First-loss	Count mobilization not disbursement; publish GEMs

▸Three actions, in priority order

For private allocators — buy the wrapper, not the project. The near-term opportunity is rated senior and mezzanine tranches of MDB originate-to-distribute programs, not direct single-project EMDE risk. Let the MDB originate and retain the junior slice; you take the de-risked, rated exposure your mandate permits.
For SWFs and PE/infrastructure funds — fill the equity gap. Target mezzanine, anchor and catalytic-equity roles, with the ~$215bn annual clean-energy equity gap as the headline opportunity. This is where return potential and additionality coincide.
For governments and DFIs — the highest-leverage public action is information and standardization. Full GEMs disclosure plus harmonized, replicable documentation does more to unlock trillions than another round of summit pledges. Re-engineer internal incentives to count mobilization, not disbursement.

The One-Line Conclusion

The $7 trillion is not missing — it is mis-wrapped. Replace the bespoke, unrated, illiquid pipe with rated, tranched, standardized capital-markets plumbing, supply the catalytic first-loss that makes senior tranches safe, and publish the data that proves EMDE risk is lower than priced. Do those three things and billions become trillions. Do none, and the paradox persists.

Gravitywell Research · Confidential34 / 47

Part V · The Private-Capital Desk

Section 09

What You Can
Actually Buy

The report so far diagnosed the market. This part is the trade desk: what is raising now, who mobilized most, what the structures really cost, how you get your money back — and the case against the whole enterprise. Decision-ready, dated to May 2026.

Reading

Pages 35 — 44

As Of

May 2026

What You'll Take Away

01

Live vehicles, not theory

A scoreboard of funds open for commitment, with size, structure and entry point.

02

The economics, decoded

First-loss, fees, FX and domicile — what each layer of the stack costs and earns.

03

The honest exit

How blended capital gets liquid — and the bear case that says it never scales.

09 — The Private-Capital DeskThe Month in Blended Finance

Current Intelligence · May 2026

09.1The Month in Blended Finance

A dated market snapshot — §§01–08 are evergreen analysis; this page and the desk that follows are current as of May 2026 and refresh each edition. What moved in May: the regulator turned its gaze to private credit, the money kept rotating out of saturated US direct lending toward emerging-market development debt, and the post-Sevilla machinery began to ship product.

$18.7bn

2025 blended commitments — record (Convergence)

73%

of 2014–24 commitments still from public sources

200–400bps

EM private-credit pickup over comparable public debt

6 May'26

FSB report flags private-credit vulnerabilities

▸The signal

On 6 May 2026 the Financial Stability Board published Vulnerabilities in Private Credit, formally putting the $1.7tn-and-growing private-credit complex under the macro-prudential lens. The sub-text for this desk: as cracks appear in saturated US direct lending, allocators are re-discovering emerging-market development debt — senior-secured, DFI-originated, and historically lower-loss than its pricing implies (Alternative Credit Investor, Mar 2026; Ninety One, “Hidden GEMs”). The diversification case, not the impact case, is doing the selling.

▸Post-Sevilla, the plumbing ships

The FfD4 “Sevilla Platform for Action” (130 initiatives, Jul 2025) is converting to product: the IDB’s FX EDGE local-currency toolbox is live, a six-government blended platform (Canada, Denmark, France, Germany, South Africa, UK) is standing up, and the $125bn Tropical Forest Forever Facility seated its governance bodies with Norway joining Brazil. The 2026 FfD Forum (New York, 20–24 Apr) set the accountability calendar.

May 2026 — The Tape

Regulation	FSB private-credit report (6 May) — EM dev-debt reframed as diversifier
FX	TCX EU Market Creation Facility live — 2–5% discount on the dollar leg
Structure	Guarantees overtake concessional lending as 2026’s dominant tool (Convergence)
Flows	US private-credit saturation pushes mandates toward EMDE senior debt

Desk View

The macro accident of 2026 — a crowded, newly-scrutinized US private-credit market — is the best demand-side catalyst blended finance has had in a decade. The capital is looking for the exact risk EMDE development debt supplies. The constraint remains the wrapper, not the will.

Sources: FSB, Vulnerabilities in Private Credit (6 May 2026); Convergence, State of Blended Finance 2025 & Four Trends 2026; Mazzucato & Vieira de Sá (UCL IIPP, Jul 2025) — public/private split; Alternative Credit Investor (Mar 2026); FfD4 Sevilla Platform for Action.

Gravitywell Research · Confidential35 / 47

09 — The Private-Capital DeskCapital In Market

The Fundraising Scoreboard

09.2Capital In Market

Theory does not accept commitments. These are the vehicles an institutional allocator could actually underwrite into now — by size, structure, status and entry point.

Table 9.1 — Selected blended / EMDE-development vehicles, 2024–2026
Vehicle	Sponsor / Manager	Size / Target	Structure	Where you enter	Status
FAST-P — Green Investments Partnership	MAS (Singapore) / IFC, FMO, BII, HSBC, Temasek	$510m close · $5bn ambition	Concessional + commercial thematic funds	Senior & mezz commercial tranche	1st close 2025
ILX Fund II	ILX (Cardano Development), Amsterdam	$1.7bn AuM (I+II)	DFI co-investment private debt (A/B loans)	Senior EMDE debt, pension-grade	Open / deploying
BII Climate Initiative	British International Investment	$1.5bn (5-yr plan)	Catalytic equity + crowd-in	Co-invest / anchor equity	Launching
Climate Finance Partnership	BlackRock + AfD, KfW, JBIC	Scaling	First-loss public layer, commercial equity	Mezz / equity	Deploying
Tropical Forest Forever Facility	Brazil + Norway (sovereign-led)	$125bn target	Sovereign-anchored results facility	Senior notes (forming)	Standing up
SDG Loan Fund	FMO / AllianzGI	$1.1bn	First-loss + guarantee, 9:1	Class A senior (closed)	Reference — §09.3

Sizes are target or latest-close figures, not committed deployment. Sources: Convergence (2025–26); Allianz Global Investors / FMO; ILX Fund / Cardano Development; British International Investment; BlackRock; TFFF / Government of Brazil.

Read The Scoreboard This Way

Two products dominate the investable surface today: DFI co-investment debt funds (ILX, FAST-P senior) for fiduciary-constrained capital, and catalytic equity vehicles (BII, CFP) for those who can take the mezzanine and equity gap. The pension-grade senior debt route is the most replicable — APG anchored ILX Fund I with $750m precisely because it looked like the EM corporate debt it already owns.

What To Demand Before You Sign

External rating or a credible shadow rating on your tranche.
DFI co-lending and preferred-creditor benefit explicit in docs.
GEMs-grade loss data for the strategy, not a glossy deck.
A defined refinancing / securitization exit (see §09.7).

Gravitywell Research · Confidential36 / 47

09 — The Private-Capital DeskThe Diligence Checklist

Before You Commit

09.3The Diligence Checklist

§09.2 listed what to demand; this is how a committee verifies it — the pass/fail tests that separate a genuinely de-risked tranche from a glossy one. Benchmark each against the SDG Loan Fund template decoded in §03.2.

Table 9.2 — The blended-finance diligence checklist
Test	What to verify	Pass	Red flag
Rating	External or credible shadow rating on your tranche	IG senior, documented	Unrated, sold as "de-risked"
Loss data	GEMs-grade, strategy-specific loss history	Pool default/recovery series	Glossy deck, no numbers
Preferred creditor	DFI co-lending explicit in the docs	Pari passu with an MDB	"DFI-aligned," no co-lend
First-loss adequacy	Subordination vs. modeled stress loss	Buffer > tail loss (cf. §05.2)	Thin junior — you are first-loss
FX	Hedge cost modeled, not assumed away	Priced (TCX) or local-ccy	Unhedged hard-ccy on local revenue
Additionality	Minimum-concessionality, sized first-loss	Explicit, priced catalytic layer	Rate subsidy buried in the coupon
Exit	Defined refinancing / securitization path	Route per §09.7	Hold-to-maturity, no secondary
Manager & fees	Origination + workout record; fee load	Cycle-tested; ≤1.5% + low carry	First-time GP; 2/20 on credit

The Five Deal-Killers

In order of how often they sink a commitment: (1) an unrated senior tranche your fiduciary screen rejects; (2) an FX mismatch nobody priced; (3) a first-loss layer too thin for the strategy's real loss distribution; (4) no preferred-creditor benefit despite "DFI" branding; (5) no exit — a long lock with no refinancing or secondary. Any one is disqualifying until cured.

The One Test That Matters

Strip the impact narrative and ask: would my committee underwrite this credit at this spread if it carried no "blended" or "SDG" label? If yes, the structure has done its job — it converted a development asset into something your mandate already buys. If the label is doing the work, pass. The SDG Loan Fund (§03.2) clears this test; most pitches do not.

Gravitywell Research · Confidential37 / 47

09 — The Private-Capital DeskThe Economics

Fees, First-Loss & Concessionality

09.4The Economics of the Stack

Private capital is bought and sold on terms. Here is what each layer of a blended structure costs, who pays it, and the single ratio that decides whether a deal is efficient public spending or a giveaway. (The cost-effectiveness ranking of instruments sits in §03.4; this is the deal-level term sheet.)

Table 9.3 — Economics by layer (indicative, EMDE blended debt fund)
Layer	Typical size	Return / cost	Who bears it	The point of tension
Senior / preference	80–90%	Market-rate (EM debt + 200–400bps)	Private investor earns	Must be rateable to count for fiduciaries
Mezzanine	5–15%	Equity-like, 2nd loss	SWF / PE / family office	Thin supply — the real gap
First-loss (junior)	~10%	Concessional / below-market	DFI / donor	Minimum-concessionality discipline
Guarantee	2–5% of senior	Fee or unfunded	Foundation / DFI	Highest leverage per public dollar
Management fee	—	~0.5–1.5% p.a.	Private investor pays	Below mainstream PE; thin economics for GPs
Carry	—	0–10%	GP	Often waived/reduced in concessional vehicles

The Leverage Math That Matters

In the SDG Loan Fund, a $25m guarantee sits beneath $1,000m of senior — a 40× notional backstop ratio on the philanthropic dollar, and a 9:1 mobilization ratio overall. This is why OECD (2026) argues fund structure, not headline concessionality, is the real driver: a small, well-placed junior layer mobilizes far more than a large, badly-placed subsidy.

The Re-Rating Catch

Amundi (Oct 2025) is blunt: an unrated senior tranche cannot re-rate, so it cannot be sold to the rated-paper buyers who hold the trillions. Concretely: under Solvency II's standard formula a 5-yr AAA senior STS securitization now carries a spread charge of only ~0.7%/yr of duration (post the Oct-2025 reform, effective Jan 2027 — near covered-bond territory), while unrated or non-senior positions run multiples higher. The rating is what moves the asset from punitive to portfolio-eligible — which is why the cost of the rating, and the GEMs data to earn it (§04), is the most valuable line item in the stack.

Gravitywell Research · Confidential38 / 47

09 — The Private-Capital DeskCurrency, Tax & Domicile

The Plumbing Nobody Pitches

09.5Currency, Tax & Domicile

Two deals die in the back office for every one that dies on credit: an un-hedgeable currency, a fund domiciled where your committee can’t invest. The unglamorous layer that decides whether capital can actually move. (The structural FX problem is framed in §02.4; here it is a priced, solvable line item.)

▸The currency problem — and the new fix

EMDE assets earn in local currency; institutional capital is priced in dollars and euros. TCX — the development-backed hedging fund — now covers 140+ frontier and emerging currencies and had enabled over $10bn of local-currency funding by 2024. The cost has been the barrier. In 2025 the EU Market Creation Facility (Pricing Component Plus), an up-to-€150m hedge guarantee run by TCX, began offering 2–5% discounts on the dollar leg — material on long-tenor frontier loans — with €170m of commitments expected to mobilize ~€2bn of local-currency finance. The IDB’s FX EDGE toolbox (launched at FfD4) extends the same logic across MDBs.

▸The takeaway

FX is no longer a reason not to do the deal — it is a line item with a shrinking, increasingly subsidized price. Hedge cost should be modeled, not feared.

Domicile — Where The Vehicle Lives

Domicile	Why it’s used
Luxembourg (SICAV-SIF / RAIF)	Fiduciary-familiar; SDG Loan Fund’s wrapper; EU passport
Cayman	Default for US / global LPs; tax-neutral
GIFT City (IFSCA FME)	India-onshore offshore; Family Investment Funds; treaty access
Netherlands / Amsterdam	DFI-adjacent managers (FMO, ILX, Cardano)

Tax & Structure Notes

Concessional returns and DFI co-investment can complicate withholding and treaty positions; the wrapper’s domicile, not the asset’s, usually governs. Rule of thumb: pick the domicile your largest anchor LP’s committee already approves, then engineer the assets in.

Sources: TCX Fund (2024–25); GlobalCapital — EU Market Creation Facility; OECD, Unlocking Local-Currency Financing in EMDEs (2025); IDB FX EDGE (FfD4, 2025); IFSCA FME framework.

Gravitywell Research · Confidential39 / 47

09 — The Private-Capital DeskThe Mobilization League

Track Records, Not Promises

09.6The Mobilization League

Allocators back managers, not asset classes. Who actually moved private money — at the system level and at the vehicle level — and how to read the numbers without being flattered by them. (This is the track-record cut; the structure registry is §05.3, the sector case studies §07.)

$87.9bn

MDB/DFI private mobilization 2023 (+24% YoY)

$14.6bn

to least-developed countries (+55%, record)

$213bn

cumulative blended deals to mid-2025 (Convergence)

4 : 1

median mobilization ratio, all deals

Table 9.4 — Mobilization track record, by vehicle (illustrative leaderboard)
Vehicle / program	Manager	Scale	Leverage	What it proves
MCPP (Managed Co-Lending Portfolio)	IFC	$16–25.5bn	High	Syndication at institutional scale
SDG Loan Fund	FMO / AllianzGI	$1.1bn	9 : 1	Guarantee-led mobilization works
ILX Fund I & II	ILX / Cardano	$1.7bn	Pension-led	APG/Danish pensions buy EMDE debt
Room2Run / Scaling4Impact	AfDB / IDB	Portfolio	Synthetic	Balance-sheet risk transfer at MDBs
$1bn+ climate deals	DFI-arranged	—	7.6×	Leverage scales sharply with size

Read The League With Suspicion

Reported mobilization mixes Private Direct (active arranging) and Private Indirect (parallel / anchor) finance, and attributes PDM at full value to the arranger — which both flatters individual institutions and risks double-counting. Treat headline league figures as an upper bound on activity, not audited additional capital (§Methodology).

The One Number To Anchor On

Leverage is not a constant — it is a function of size. DFI-arranged climate deals above $1bn mobilize 7.6×; sub-scale deals barely clear 2×. The implication for an allocator: write fewer, larger tickets into bigger pooled vehicles, and let the arranger aggregate the small stuff.

Gravitywell Research · Confidential40 / 47

09 — The Private-Capital DeskLiquidity & Exits

How You Get Your Money Back

09.7Liquidity, Exits & Secondaries

The question every committee asks last and means first. Blended finance has historically been a roach motel for capital — money checks in, distributions don’t check out. That is finally changing, and the mechanism is the report’s whole thesis.

▸The exit IS the re-plumbing

For blended debt, the cleanest exit is not a sale — it is a refinancing into the capital markets. Originate-to-distribute, securitization and true-sale CLOs (§05) are not just origination tools; they are the liquidity event. The DFI warehouses and seasons the asset, then sells rated paper to the senior buyer. The wrapper that makes it bankable is the same wrapper that makes it exitable.

▸The secondary market arrives

For equity and fund stakes, the private-markets liquidity revolution is now reaching EMDE impact. GP-led secondary volume hit $115bn in 2025; continuation vehicles were 89% of GP-led and ~43% of all secondary volume, with 46% of GPs in the 2026 outlook using them to manufacture DPI. LP-led stakes are clearing near 89% of NAV. EMDE impact secondaries remain nascent but are following the same path.

The Three Exit Routes

1 · Refinance / securitize — sell rated senior paper to capital markets. The primary blended-debt exit.

2 · GP-led / continuation — manufacture DPI without an asset sale. New to EMDE impact.

3 · DFI step-back / refinance — development lender refinances out the early commercial investor.

What To Demand

A defined liquidity path, in the docs, before commitment: target refinancing window, securitization eligibility, and whether the DFI will warehouse. “Hold to maturity” is a strategy only if maturity is short and certain.

Sources: Dechert / White & Case / Hamilton Lane — GP-led & continuation-vehicle data (2025); 2026 Global Private Equity Outlook; Delphos — EM secondaries (2025); §05 (this report) — securitization mechanics.

Gravitywell Research · Confidential41 / 47

09 — The Private-Capital DeskThe Family Office Angle

Patient Capital, Fewer Rules

09.8The Family Office Angle

The investor type the playbook usually forgets — and the one structurally best-suited to the gap. Family offices carry no Solvency II, no Basel, no quarterly-mark tyranny, and can blend their own philanthropy with their own balance sheet in a single decision.

~45%

of family offices weigh sustainability (UBS GFO 2025, n=317)

~33%

globally engaged in responsible investing (Campden)

~25%

North America — the laggard, the upside

$25m

a foundation guarantee that moved $1bn (§09.3)

▸The structural edge

Where pensions and insurers are forced into the rated senior tranche, a family office can sit anywhere in the stack — mezzanine, anchor equity, even the catalytic first-loss — because no prudential regulator dictates its risk slice. It is the natural buyer of the ~$215bn EMDE clean-energy equity gap that institutional mandates cannot reach.

▸The MacArthur move

The single most leveraged thing private wealth can do is what the MacArthur Foundation did: write the unfunded guarantee that makes someone else’s senior tranche bankable. A family with both a foundation and an investment office can supply the catalytic layer and co-invest the commercial layer — capturing impact and return in one structure.

The Family Office Playbook

Anchor the equity gap — mezz & catalytic equity others can’t hold.
Deploy the foundation — guarantees & first-loss from the philanthropic side.
Use GIFT City / Lux FIFs — co-investment wrappers built for private wealth.
Co-invest beside DFIs — borrow their origination and preferred-creditor status.

Why It Matters

The capital the market most needs — patient, flexible, junior-tolerant — is exactly the capital family offices hold and institutions don’t. North America’s low ~25% engagement is not a warning; it is the headroom.

Sources: UBS Global Family Office Report 2025 (n=317); Campden Wealth / RBC North America Family Office Report 2025; SDG Loan Fund disclosures; §07 — clean-energy equity gap.

Gravitywell Research · Confidential42 / 47

09 — The Private-Capital DeskBlended vs The Book

Cross-Strategy Benchmarking

09.9Blended vs The Book

No allocator buys blended finance in isolation. They buy it instead of — or alongside — the private credit, infra debt and EM corporate exposure they already own. Here is how the trade compares on the only axes that decide the allocation.

Table 9.5 — EMDE development debt vs comparable strategies (indicative, 2025–26)
Strategy	Yield / return	Realized loss	Liquidity	Mandate fit
EMDE development debt (senior)	EM debt +200–400bps	3.54% default / 72.9% recovery (GEMs)	Low → improving (§09.7)	Fiduciary if rated
US private credit / direct lending	~SOFR +500–600bps	Rising; newly FSB-scrutinized	Low	Crowded, saturated
Infrastructure debt (BB)	~7%+ (SOFR +high-200/300s)	Low, long-tenor	Low	Insurer-friendly
EM corporate debt (public)	CEMBI-grade	Market, mark-to-market	High	Liquid sleeve
The blended edge	≈ IG loss, EM-debt yield	Senior-secured position + DFI preferred-creditor benefit + diversification from US credit

The Portfolio Case

GEMs data show EMDE private default at 3.54% with 72.9% recovery — sub-Saharan defaults priced like S&P single-B yet recovering above 78%. That is an investment-grade loss profile attached to an emerging-market yield. As US private-credit risk concentrates (FSB, May 2026), EMDE development debt is one of the few credit sleeves offering genuine diversification rather than more of the same beta.

The Honest Caveat

GEMs statistics reflect MDB/DFI experience — preferred-creditor treatment and intensive workout — and should be haircut for non-MDB lenders. The diversification is real; the loss data does not transfer one-for-one to a fund without DFI co-lending. Buy the structure that buys you the preferred-creditor benefit, not just the asset.

Gravitywell Research · Confidential43 / 47

09 — The Private-Capital DeskThe Bear Case

The Opposing View · Signed

09.10The Bear Case

A magazine that only prints its own thesis is a brochure. Here is the most serious intellectual case against everything this report recommends — stated at full strength, followed by our answer. It extends the additionality evidence and the Gabor "de-risking state" critique already laid out analytically in §06.1; this is the signed, investor-facing version.

▸The prosecution: “Mind the Mission, Not the Gap”

Mariana Mazzucato and Rogério Vieira de Sá (UCL IIPP, Jul 2025) argue the entire model is mis-specified. Between 2014 and 2024, ~73% of blended-finance commitments still came from public sources and only ~27% from private actors — the “billions to trillions” promise simply did not arrive. Worse, blended finance socializes risk and privatizes reward: public money de-risks the deals private capital would arguably have done anyway, while flowing to the lowest-risk sectors and geographies, leaving the hardest, most additional projects unfunded. The fix is not better plumbing — it is mission-oriented public finance that directs capital, rather than subsidizing it to show up.

Private initiative reorients public resources toward lower-risk, higher-return investments — the opposite of what development finance is for.Mazzucato & Vieira de Sá, UCL IIPP, 2025

The Defense

The critique is right about the past and wrong about the trajectory. Yes — bespoke, unrated, sub-scale blended deals mobilized little and chased safe assets. That is precisely the “mis-wrapped” market this report indicts. The 73/27 split is an indictment of structure, not of private capital’s willingness: where instruments became rated, standardized and liquid (MCPP, the SDG Loan Fund, GEMs-backed securitization), the ratios inverted toward 7–9:1.

Where We Concede

Mazzucato wins one point outright: additionality must be proven, not assumed. If public money only de-risks deals that would have happened, it is a subsidy, not a catalyst. Our answer — minimum-concessionality discipline, GEMs disclosure, and counting mobilization not disbursement — is an attempt to earn the additionality the critique rightly demands.

Source: Mazzucato, M. & Vieira de Sá, R. (2025), Mind the Mission, Not the Gap: Rethinking blended finance for public purpose, UCL Institute for Innovation and Public Purpose Working Paper (Jun–Jul 2025).

Gravitywell Research · Confidential44 / 47

AppendixMethodology & Sources

For The Technical Reader

APPENDIXMethodology, Caveats & Sources

Measurement is contested

Three regimes — OECD, the DFI/MDB Joint Working Group, and Convergence — define "blended finance" and "mobilization" differently, producing non-comparable totals. The MDB methodology separates Private Direct Mobilization (active arranging) from Private Indirect Mobilization (anchor/parallel finance), attributing PDM at full value to the arranger, which both flatters individual institutions and risks double-counting. Publish What You Fund's What Works (Oct 2024) argues the methodology understates emerging forms (portfolio risk transfers, balance-sheet operations) while lacking the project-level disclosure needed to verify additionality. Reported mobilization is best read as a lower bound on activity but an upper bound on genuinely additional private capital.

Figures are not additive

Headline figures (~$18bn 2024 blended market and 1,350-deal/$249bn cumulative database, Convergence 2025; ~$70bn OECD/MDB mobilization; ~$88bn 2023 MDB total private mobilization; ~$4tn SDG gap; ~$10tn cumulative shortfall to 2050 per OECD) come from differing methodologies and base years and must not be summed.

Self-reported & forward-looking flagged

The >$650bn decade CAF headroom (MDB Heads, Jun 2025), the MIGA Platform's $12.3bn/$20bn-by-2030 figures, IFC local-currency targets and Mission 300 are institutional projections or self-reported results, not independently audited outcomes. The AfDB multi-originator SST (Dec 2024) was a Letter of Intent to explore a structure, not a priced deal. Tokenization/digital-MRV claims rest on a thin, largely promotional evidence base and are treated cautiously.

Structural distinctions

Room2Run and IDB Scaling4Impact are synthetic securitizations (risk transferred, assets retained); the IFC 2025 deal is a true-sale cash CLO. GEMs default/recovery statistics reflect the specific experience of MDBs/DFIs (with preferred-creditor treatment) and may not generalize.

Conflicting datapoints disclosed in-text

Divergent MCPP totals ($16bn / $19bn / $25.5bn across dates) and CAF headroom ranges ($60–572bn bottom-up; ~$300–400bn G20 2025 near-term CAF; >$650bn total decade headroom, MDB Heads Jun 2025). Where sources conflict, the most recent primary-source figure is preferred and the range disclosed. The $600–800bn "data dividend" is an estimate by the Center for Global Development of a contingent re-pricing opportunity, not an audited market value, and is realized only if rating agencies and prudential regulators respond to disclosure.

Returns & loss data

Return ranges (§04.1) blend the only audited pooled benchmark (Cambridge/GIIN, ~6.9% IRR — 2015-vintage, dated) with current self-reported returns (GIIN State of the Market 2025; ~16–18% EM/DM PE per 2024 ed.) carrying survivorship bias; we treat the pooled figure as a floor, survey data as a ceiling. GEMs default/recovery statistics (§04.2) reflect the MDB/DFI experience — preferred-creditor treatment, intensive workout — and should be haircut for non-MDB lenders. Failure cases (§04.4) are illustrative of documented failure modes, not an exhaustive loss census.

Gravitywell Research · Confidential45 / 47

AppendixSource Register

Citations

APPENDIX · iiSource Register

Superscript numerals throughout the report refer to the primary sources below. Figures were verified against the most recent primary-source publication available at the time of writing (June 2026).

1 Convergence, State of Blended Finance 2024 (Climate Edition); TodayESG summary — leverage by deal type/size; $500m+ climate deals with DFIs mobilize 5.9x.

2 Convergence, "Blended finance leverage ratio" (Market Insights) — leverage rises with transaction size to 7.6x at $1bn+.

3 Convergence, State of Blended Finance 2025 (Spring) — 123 deals / $18bn in 2024; database 1,350 deals / $249bn; median deal $65m; grants 41%→10% of disbursement (2022–24).

4 GEMs Consortium / EBRD, EIB (Oct 2025) — private default 3.54%, recovery 72.9%; financials 2.26%/79.1%; SSA 6.05%/>78%; sovereign 0.77%/95.1%.

5 IFC / GEMs (Oct 2024) — sovereign-guaranteed default 1.06%, recovery 94.9%.

6 Center for Global Development, GEMs and the $600–800 Billion Dollar Data Dividend (Dec 2025) — private median recovery >90%; SSA default ≈ S&P B/B−.

7 EBRD / GEMs (Oct 2024) — private default 3.56%, recovery 72.2% (prior release).

8 GEMs Consortium, Research Note (2025) — default by income level; low-income 6.3%; high-income 2.3%.

9 Cambridge Associates / GIIN, Impact Investing Benchmark (2015 vintage; pre-2011 funds — dated, the only audited pooled series) — pooled IRR 6.9% vs. 8.1% comparator; EM 6.2%; Africa 9.7%.

10 ImpactAlpha / GIIN — benchmark interpretation; dispersion > average.

11 Amundi Research Center (Oct 2025), Demystifying Credit Enhancements — unrated tranches cannot re-rate; mobilization limits.

12 GIIN, State of the Market 2025 (Oct 2025; n=429, 54 countries) — 90% of investors met or exceeded financial-return expectations; 79% target risk-adjusted market-rate returns; impact AUM ~$1.6tn. Realized PE returns ~16% DM / ~18% EM per the 2024 edition. Survey self-reports, not a pooled IRR.

13 SRI360 / sector reporting — EM utility & off-grid energy IRRs (8–9%) and write-downs.

14 J.P. Morgan / GIIN, Impact Investments: An Emerging Asset Class — CEMBI & DM HY benchmarks.

15 Dimos & Pugh, via OECD Development Co-operation Working Paper 91 (2021) — rejects full crowding-out; no substantial additionality.

16 OECD (2026), "Making blended finance work: why fund structure is the real driver" — junior tranches; minimum concessionality.

17 IFC / DFI Working Group, enhanced additionality guidelines (2025); Sustainability Atlas summary.

18 OECD, DAC Blended Finance Guidance 2025 — "cottage industry"; GEMs as evidence risk is lower than perceived.

19 BII investor-contribution methodology, via OECD blended-finance case studies (2025).

20 Mazzucato et al. (2025), Mind the Mission, Not the Gap — structural critique of financial engineering. Daniela Gabor, "The Wall Street Consensus," Development and Change 52(3), 2021 — the "de-risking state" critique.

21 G20 Roadmap for MDB CAF Implementation (GI Hub) — ~$200bn headroom/decade; callable capital unimplemented; GEMS 2.0 support.

22 CGD / MDB Reform Accelerator — CRA willingness to revisit callable-capital methodology.

23 MDB Heads Joint Statement (28 Jun 2025) — >$650bn total lending headroom over the decade (>$250bn added since mid-2024); hybrid & risk-transfer instruments; callable-capital clarity. Plus FfD4 "Sevilla Commitment" & Sevilla Platform for Action (Jun–Jul 2025).

24 IFC / World Bank, EM Securitization 2025-1 (closed 19 Sep 2025; $510m; senior $320m Aaa at SOFR+130, 37.3% CE) — arranger Goldman Sachs, LSE-listed. Basel securitization framework (SEC-SA/SEC-IRBA; STS criteria; 5% retention).

Additional: GI Hub Blended Finance in Infrastructure (2024, $0.4 private per $1); IDB Invest & World Bank securitization releases; MIGA Guarantee Platform; FMO/Allianz SDG Loan Fund disclosures; CPI Global Landscape of Climate Finance 2025; UN FSDR 2026 (Apr 2026) & UNCTAD WIR 2025 SDG gap.

About Gravitywell Research

Gravitywell Research produces institutional-grade investment intelligence at the intersection of capital markets, technology and global policy. This report is for professional and institutional investors and is provided for information only. It is not investment advice, nor an offer or solicitation, and does not constitute personalized financial, legal or tax advice. Gravitywell Research is not a licensed investment or financial adviser; readers must conduct independent diligence and consult licensed advisers before acting. All figures are estimates subject to the methodology and caveats herein.

Report GWR-2026-BF-001 · Consolidated Edition · June 2026 · Institutional / For Professional Investors

Gravitywell Research · Confidential46 / 47

AppendixGlossary & Acronyms

Decoding The Jargon

APPENDIX · iiiGlossary & Acronyms

Blended finance carries dense, overlapping terminology. The working definitions below reflect usage in this report; where market usage diverges, the report text governs.

Additionality: Evidence that public/concessional support produced investment that would not otherwise have occurred — the core test of whether blending is justified or merely subsidising commercial capital.
Blended finance: The use of concessional (below-market) public or philanthropic capital to mobilise commercial private capital into development and climate projects.
Callable capital: Shareholder capital pledged to an MDB but not paid in — a contingent backstop (~$1.2tn system-wide) that underwrites the AAA ratings but is almost never called.
Catalytic / first-loss capital: Junior capital that absorbs the first dollar of loss, making the senior tranche safe enough for institutional investors. The scarcest input in the system.
Concessional capital: Funding provided below market terms (lower rate, longer tenor, or subordinated) to improve a deal's blended economics.
Credit enhancement: Any mechanism — guarantee, insurance, subordination — that improves a security's risk profile so it can attract regulated capital or a credit rating.
GEMs: Global Emerging Markets Risk Database — a consortium of MDBs/DFIs pooling default and recovery statistics. Its disclosure underpins the potential re-rating of EM credit risk.
Leverage / mobilisation ratio: Private capital catalysed per unit of public/concessional input. System-wide it has been near 1.8x; it rises with deal size to ~7.6x on $1bn+ transactions.
Originate-to-distribute (O2D): A model in which an originator (e.g. an MDB) makes loans, then sells or securitises them to investors — recycling balance sheet rather than holding to maturity.
Preferred-creditor treatment: The de facto seniority MDBs enjoy in repayment, a key reason GEMs default/recovery data may not generalise to private lenders.
Private mobilisation (PDM / PIM): Private Direct Mobilisation (active arranging, attributed in full to the arranger) vs Private Indirect Mobilisation (anchor/parallel finance). The split drives much of the measurement dispute.
Securitisation (synthetic vs true-sale): Pooling assets into tradable tranches. Synthetic transfers risk while retaining the assets (e.g. Room2Run); true-sale sells the assets outright (e.g. the 2025 IFC cash CLO).
Subordination / tranching: Slicing a structure into junior, mezzanine and senior layers so that losses hit the junior tranche first — the engineering that lets one risk profile serve many investor types.
Technical assistance (TA): Grant-funded support that builds the project pipeline and local capacity — the enabling input behind most bankable deals, but grant-dependent and ODA-exposed.

Institutional Abbreviations

AUMAssets under management BRIBelt and Road Initiative (China) CAFCapital Adequacy Frameworks (G20 MDB review) CLOCollateralised loan obligation DACDevelopment Assistance Committee (OECD) DFIDevelopment finance institution EMDEEmerging-market & developing economies FfDFinancing for Development (UN process) GFANZGlasgow Financial Alliance for Net Zero IRRInternal rate of return MDBMultilateral development bank MIGAMultilateral Investment Guarantee Agency (WBG) ODAOfficial development assistance PRIPolitical risk insurance SDGSustainable Development Goal(s) SDRSpecial Drawing Rights (IMF) SSTSynthetic securitisation transaction TCXThe Currency Exchange Fund

Gravitywell Research · Confidential47 / 47

Gravitywell ResearchInstitutional Investment Intelligence

End of Report

Mis-wrapped,
not missing.

The capital exists. The appetite exists. What is missing is the standardized, rated, liquid wrapper — and the data to prove the risk is lower than priced. Build those, and the paradox dissolves.

Gravitywell Research · GWR-2026-BF-001 · June 2026 Confidential · For Professional Investors

© 2026 Gravitywell Research. For information only; not investment, legal or tax advice and not an offer or solicitation. Figures are estimates subject to the methodology and caveats herein. Gravitywell Research is not a licensed investment adviser. Past performance and modeled scenarios are not indicative of future results.

The $7 TrillionParadox.

CONTENTSNavigating the Deployment Gap

00A 1–2% Solution to a Trillion-Dollar Problem

00 · iiSix Findings That Reframe the Debate

It is a pipeline problem, not a capital problem.

The ratio isn't frozen — it's fragmented.

Basel, Solvency II and fiduciary rules penalize the wrapper.

This is a fixed-income-like asset class, not a venture one.

Disclosure beats another summit pledge.

Some "mobilization" isn't additional.

00 · iiiWhat To Do With This

Buy the wrapper, not the project.

Fill the equity gap.

Standardize and disclose.

QUICK READFive Calls, Five Risks

Anatomy ofthe Gap

01.1Committed Is Not Deployed

▸The three measurement regimes

01.2Leverage Isn't Frozen — It's Fragmented

01.3The Geography of Under-Deployment

The concentration problem

▸Sectoral concentration mirrors the geography

The Barriers

02.1Six Constraints, One Broken Pipe

① Pipeline — the binding constraint

② Regulation — Basel, Solvency II, fiduciary

③ Currency risk

④ ESG & taxonomic fragmentation

⑤ Transaction costs & non-standardization

⑥ Political & ODA retrenchment

02.2Regulation Rejects the Wrapper, Not the Credit

▸The GEMs evidence: perception vs. reality

▸How the rules bite

02.3Severity vs. Tractability

High severity, low tractability

High severity, high tractability

02.4Currency, Macro & China

▸Currency: the structural killer

▸The China question

InstrumentEffectiveness

03.1The Instrument Effectiveness Index

03.2The SDG Loan Fund — 9:1, Decoded

Why it worked

Why it is hard to replicate

03.3Where the Mechanisms Disappoint

Green bonds in frontier markets

Sustainability-linked debt

PPPs at scale

▸The additionality question hangs over everything

03.4Where Scarce Public Money Works Hardest

Returns, Losses& the Re-Rating

04.1The Return Profile, Honestly

04.2The Loss Record, Decomposed

Three things the data settle

04.3The Data Dividend, Derived

04.4What Failure Looks Like

The unrated-tranche trap

Off-grid energy write-downs

The impact-return drag

PPP contingent liabilities

The Capital-MarketsRe-Plumbing

05.1From Balance-Sheet Lending to Originate-to-Distribute

05.2Tranching, True-Sale and Synthetic Risk Transfer

True-sale (cash) securitization

Synthetic risk transfer (SRT)

▸Worked example — IFC's inaugural CLO (2025), and how the senior stays safe

05.3Registry of Landmark Structures

05.4From Pilot to Platform

What's still missing

Counterfactual& Critique

06.1Is Any of This Additional?

06.2The Borrower's-Eye View

What the borrower actually experiences

Transaction Registry& Case Studies

07.1Where Renewables Became Bankable

Scaling Solar

Climate Investor One

07.2Healthcare, Inclusion & Infrastructure

IFFIm

Health Finance Coalition

The $7 Trillion
Paradox.

Anatomy of
the Gap

Instrument
Effectiveness

Returns, Losses
& the Re-Rating

The Capital-Markets
Re-Plumbing

Counterfactual
& Critique

Transaction Registry
& Case Studies

Outlook &
Investment Implications

What You Can
Actually Buy

Mis-wrapped,
not missing.