A− A+ Close ×

THE DEVELOPMENTAL CAPITAL VELOCITY FRAMEWORK

Repositioning Securitisation as an Instrument of Capital Recycling for African Development Finance Institutions

Prince Sarpong, PhD

Sarpong Strategia

prince@sarpongstrategia.com | www.sarpongstrategia.com

Working Paper. July 2026.

Abstract

Development finance institutions measure success by capital deployed and portfolio quality, metrics that reward holding assets. This paper argues that a more consequential dimension has been overlooked: the velocity with which scarce public capital can be recycled into successive developmental investments without compromising prudential stability. The argument is urgent: official development assistance recorded its steepest contraction on record in 2025, with grants falling fastest, so capital that cannot be replenished must be recycled. The paper restates the institution’s task as a constrained optimisation, with velocity as the control variable, measured by an annual velocity ratio and a cumulative capital multiplier computed on capital genuinely freed rather than on assets placed. Velocity is raised by recombining existing instruments: stratified securitisation of a seasoned pool, credit enhancement that makes the senior tranche eligible for domestic pension capital, and a binding reinvestment covenant that preserves additionality. The capital analysis is conditioned on institutional type, treating rating-agency models as binding where no supervisor holds jurisdiction, and the retained first loss is split and priced explicitly, with its capital charge deducted from the release. A stylised illustration suggests the same capital could finance roughly twice the origination of a hold-to-maturity model over a decade; ignoring the retained position would claim nearly three times. The framework requires no new legislation; it extends the African Development Bank’s Room2Run securitisation and the G20 capital-adequacy agenda.

Keywords: development finance institutions; securitisation; capital velocity; credit guarantees; Basel III and IV; rating-agency capital; additionality; blended finance; pension capital.

JEL classification: G23, G28, O16, H81.

1. The strategic imperative

Development finance has traditionally measured success by two quantities: the volume of capital deployed and the quality of the portfolio that results. Both matter, and neither is sufficient. They describe a stock, the assets a development finance institution holds, and they say nothing about a flow that is at least as important for a capital-scarce continent, namely the rate at which a given unit of institutional capital can be returned to work in new developmental investments. An institution that lends one Rand once and holds the exposure for twenty years, and an institution that lends the same Rand, recovers it through disciplined recycling, and lends it again several times over, will report similar deployment and similar portfolio quality. Their developmental output is very different.

The proposition of this paper is that the objective embedded in existing measurement is incomplete and should be restated. The theory and practice of development finance institutions optimise stocks, and the literature has likewise concentrated on mandates, governance and portfolio performance rather than on the turnover of capital (Calice, 2013; Khadiagala, 2011; Léon, 2025). This paper proposes optimising turnover under constraint: the institution’s task is to maximise the cumulative developmental origination that a unit of regulatory capital can finance over a horizon, subject to acceptable risk, capital adequacy, liquidity and institutional sustainability. Developmental Capital Velocity names the rate at which capital turns under that objective, and it is introduced here as an institutional design principle rather than a financing technique. Every section that follows is evidence for the proposition rather than decoration around it: a development finance institution should be judged as much by how many times its scarce capital can be made to turn as by what it holds.

The urgency of the proposition is supplied by an external funding environment that has turned decisively against the model the old objective assumed. Official development assistance fell by six per cent in real terms in 2024 and by a further 23 per cent in 2025, to 174 billion United States dollars, the steepest contraction on record and the second consecutive annual decline; within bilateral assistance, grants fell by 29 per cent against 10 per cent for loans, and further declines are projected through 2028, with bilateral assistance to sub-Saharan Africa projected to fall by between 16 and 28 per cent in 2025 alone (OECD, 2025; OECD, 2026). The consequences for development finance institutions run through every channel that once relaxed their capital constraint. Concessional and donor envelopes for guarantees, first-loss capital and blended structures are shrinking, and shrinking fastest in exactly the grant form on which such structures depend. Sovereign shareholders, fiscally constrained and borrowing expensively, are less able to recapitalise the institutions they own. Market funding has repriced accordingly, while developmental financing needs continue to widen (UNCTAD, 2019). An institution that cannot expect fresh equity is left with slowly accumulating retained earnings, expensive and rating-constrained hybrid instruments, and the productivity of the capital it already holds; only the third lies fully within its own control. In this environment the velocity objective stops being an optimisation refinement and becomes the operating condition of continued relevance: recycling is what replenishment used to be.

The argument is timely. The G20 Independent Review of the Capital Adequacy Frameworks of multilateral development banks concluded in 2022 that existing capital could support materially greater lending if it were managed more efficiently, including through risk transfer, while preserving credit ratings (G20, 2022). The African Development Bank had already demonstrated the mechanics in 2018, when its Room2Run transaction transferred the credit risk on a one billion dollar slice of its non-sovereign portfolio to private investors and committed the freed capital to new lending (African Development Bank, 2018). Developmental Capital Velocity takes the logic of that agenda, which has been articulated largely at the level of multilateral banks, and reframes it as an operating objective for national and regional development finance institutions, with a metric attached and with additionality protected by design.

The framework requires no new legislation, no new institutions, and no radical change to existing financial markets. It recombines regulatory, capital-market and risk-sharing mechanisms that are already in use into an integrated structure whose purpose is to raise the velocity of developmental capital. The contribution is the redefinition of the objective that existing instruments are asked to serve, rather than the invention of a new instrument. The remainder of the paper defines the velocity and multiplier measures, states the institutional objective as a constrained optimisation, and is honest about what the measures do and do not capture (Section 2), sets out the structure of the framework (Section 3), prices the senior guarantee that makes placement viable (Section 4), establishes which capital regime actually binds each type of institution and prices the retained first loss under it (Section 5), positions the framework against existing practice (Section 6), confronts the risks that velocity itself introduces (Section 7), and sets out an implementation pathway and its enabling conditions (Sections 8 and 9).

2. Defining and measuring Developmental Capital Velocity

A principle that cannot be measured cannot be managed, and the first task is therefore to give velocity a definition precise enough to compute. Let a development finance institution hold usable risk capital of C, allocated to its developmental book. Let the capital intensity of origination be k, the amount of capital consumed per unit of exposure originated, which under a simple standardised treatment is the product of the capital adequacy ratio and the average risk weight of the book. For a capital adequacy requirement of 10 per cent and an average risk weight of 100 per cent, k is 0.10, so that one unit of capital supports ten units of exposure.

Under a hold-to-maturity model, the institution originates exposure of C divided by k and then holds it. Its cumulative developmental origination over any horizon is fixed at that single figure. Under the velocity framework, once a tranche of the book has seasoned, part of it is placed with external investors, releasing capital that is redeployed into new origination, which in turn seasons and is recycled again. Let θ denote the share of each seasoned pool placed with investors. It is tempting to treat θ as the release fraction. The temptation must be resisted, because the portion of each pool the institution keeps, the first-loss position of the structure Section 3 sets out, continues to consume capital after issuance, in some regimes heavily, so the capital genuinely freed is smaller than the placed share suggests. Let ρ denote the net capital release fraction: the capital freed by a completed recycling cycle divided by the capital the pool consumed before the cycle. The effective redeployment ratio per cycle, r, is the lesser of ρ and the redeployment that the reinvestment covenant’s liquidity rules permit; in the configurations examined in this paper capital is the binding constraint, so r equals ρ.

Formally, if n recycling cycles occur within the horizon, cumulative origination follows a geometric series, and its ratio to origination under the hold model is:

V = (1 − rⁿ) / (1 − r)

As the number of cycles grows, V approaches its ceiling of 1 divided by (1 − r). Two distinct measures must now be named, because they are easily and consequentially conflated. The expression above is a multiplier: it records cumulative capital productivity, the total developmental origination a unit of capital finances over a horizon, and it rises mechanically with the number of cycles, so a multiplier quoted without its horizon is meaningless. Velocity, in the economic sense, is a rate: turnover per unit of time, which does not rise merely because the observation window lengthens. The framework therefore defines and reports both. Developmental Capital Velocity is the annual rate: new developmental origination in a period divided by the average usable capital employed in it. Under a hold-to-maturity model this rate falls toward zero once the book is fully deployed; under recycling it is sustained cycle after cycle, and sustaining it is what the framework is for. The Developmental Capital Multiplier is the cumulative expression above, the accumulation of that velocity over the horizon. One measures speed, the other measures cumulative effect, and Section 8 proposes both for annual reporting, together with the net release fraction actually achieved in each cycle.

One further point of honesty about what the multiplier measures. The cumulative origination it records is funded overwhelmingly by external investors; the incremental assets sit on pension-fund and asset-manager balance sheets, and the institution’s own capital carries the retained positions and the new, unseasoned book. The multiplier is therefore a joint measure of velocity and mobilisation: it records the developmental origination that a unit of the institution’s capital makes possible, and it should be read alongside, never in place of, measures of the risk the institution itself continues to hold. Claiming the multiplier as origination financed by the institution’s own balance sheet would overstate the case, and the framework does not make that claim.

The institutional objective, stated as an optimisation

These measures serve an objective, and the objective deserves formal statement, because it answers the question a careful reader should ask: why is velocity desirable at all? The institution’s problem is a constrained optimisation. Maximise cumulative developmental impact over the planning horizon, subject to five binding constraints: a minimum capital adequacy ratio, whether statutory or self-imposed through the market discipline described in Section 5; maintenance of the credit standing on which the institution’s funding model depends; a refinancing constraint, under which no origination is committed against placements not yet completed; a liquidity floor sufficient for the committed pipeline; and the additionality covenant, which restricts redeployment to the developmental frontier. Velocity is the control variable of this problem, never its objective. Raising the redeployment ratio increases the objective only while every constraint continues to hold; past the point at which any constraint binds, further velocity subtracts from the objective rather than adding to it, which is the formal content of Section 7’s warning that speed is itself a hazard.

Each guardrail in the framework is one of these constraints made operational. The funded junior retention enforces the underwriting standard on which the rating constraint depends; the sequencing rule enforces the refinancing constraint; the liquidity buffer enforces the liquidity floor; and the reinvestment covenant enforces additionality. The framework does not claim to solve the optimisation in closed form. It claims something more modest and more useful: that the recycling structure of Sections 3 to 5 is a feasible policy that raises the objective relative to the hold-to-maturity policy while remaining inside the constraint set, and that the two measures defined above are the instruments by which a board can verify that claim year by year.

A stylised illustration

Consider an institution with one hundred million dollars of usable capital and a capital intensity of 10 per cent, so that the first cycle originates a pool of one billion dollars. On securitisation, the senior and mezzanine tranches, together 85 per cent of the pool, are placed with investors. The first-loss position of 15 per cent is retained and, following Section 5, is split: a funded junior sublayer of 5 per cent of the pool remains on the balance sheet and attracts the punitive first-loss capital treatment, while the upper 10 per cent is covered by unfunded protection from a supranational guarantor in the highest rating band, recognised at a substitution risk weight of 20 per cent. Capital consumed after issuance is therefore 50 million dollars on the junior sublayer plus 2 million on the protected sublayer, some 52 million in total, against the 100 million the pool consumed before the transaction. The net capital release fraction ρ is 0.48. Cash proceeds of 850 million, of which the reinvestment covenant of Section 3 reserves a 20 per cent liquidity buffer, comfortably exceed the 480 million of new origination the freed capital can support, so capital, and never liquidity, binds. With a cycle length of two years and a ten-year horizon, five cycles occur. The table sets out the result.

On these assumptions the same one hundred million dollars of capital finances roughly twice the developmental origination of the hold model over a decade, approaching a ceiling just under two times as cycles accumulate. Two comparisons discipline the reading of that figure. The first is with the naive computation: equating the release fraction with the placed share of 85 per cent, so that after the covenant’s 20 per cent liquidity buffer some 68 per cent of each pool is redeployed, yields a multiplier of about 2.7 and a ceiling above three. The gap between 2.7 and 1.9 is precisely the capital consumed by the retained first loss, and reporting the larger figure amounts to pretending that retention is free. The second comparison reveals where velocity actually comes from. The multiplier is most sensitive to the thickness of the funded junior sublayer, which Section 5 sizes to the pool’s expected loss subject to a retention floor that depends on the issuing jurisdiction. An institution whose loss history justifies a junior sublayer of 3 per cent achieves ρ of about 0.68 and a five-cycle multiplier of about 2.7. Underwriting quality is therefore the engine of velocity, an alignment of incentives to which Section 7 returns. The general sensitivities follow the same logic: velocity rises with the net release fraction and the number of cycles, and falls with the cost and duration of each recycling, so the entire framework can be read as an effort to raise the net release fraction and shorten the cycle without breaching prudential limits.

3. The structure of the framework

The structure raises the net release fraction and shortens the cycle through three sequenced components. Each is drawn from established practice; the novelty lies in their combination and in the covenant that governs the proceeds.

Phase 1. Stratified asset pooling

To prevent adverse selection and to protect the institution’s own credit standing, the special purpose vehicle acquires a vertical slice of the balance sheet rather than a hand-picked selection of the best or the worst assets. The pool is a representative cross-section of infrastructure debt, mid-tier corporate financing and stabilised agricultural lending. A seasoning floor applies: assets must demonstrate a minimum of eighteen to twenty-four months of unblemished payment history after any grace period or construction phase. Seasoning establishes that cash flows have stabilised without so constricting the eligible pool that issuance becomes impractical.

Phase 2. The credit-enhancement engine

The vehicle issues a local-currency collateralised loan obligation in three tranches. The denomination is deliberate: exchange-rate risk remains among the principal deterrents to long-term private participation in infrastructure across emerging markets (Global Infrastructure Facility and International Monetary Fund, 2024), and a structure funded by domestic pension money in domestic currency avoids importing it. Existing regional credit-enhancement models, such as the African Development Bank’s partial credit guarantee capability and domestic vehicles of the InfraCredit type in Nigeria, are used to lift the senior tranche to an investment-grade rating that pension funds may hold. The tranche structure is summarised below.

The design choice that distinguishes this structure from most precedents is the placement of the senior tranche, the largest and least risky block, with domestic long-term investors. It is the de-recognition of that block that releases the most capital, and it is the guarantee on that block that makes the release feasible under pension regulation.

Phase 3. The reinvestment covenant

Liquidity released by the securitisation is governed by a binding mandate so that the institution does not drift into the role of a commercial asset manager. A liquidity buffer of 20 per cent of proceeds is placed in short-term, highly liquid sovereign paper to manage the institution’s own asset and liability parameters and to absorb slippage in issuance timing. The developmental frontier, the remaining 80 per cent, is mandated for deployment into early-stage, high-impact origination such as greenfield infrastructure, which preserves the institution’s principle of additionality. Two constraints govern redeployment, and the binding one is whichever is smaller: the covenant governs cash, while capital adequacy governs origination capacity, and in the configurations of this paper capacity binds. The covenant is the instrument that converts a balance-sheet transaction into a developmental one, and it is the reason the framework can claim to raise developmental throughput rather than merely financial efficiency.

4. Pricing the senior guarantee

A guarantee makes the senior tranche eligible for pension capital, but a guarantee also has a price, and the central friction in this kind of structure is that the premium can consume the very yield spread that is meant to attract the investor, driving the net return below the sovereign risk-free rate. The pricing of a standard credit guarantee can be written as the sum of three components:

Premium = (PD × LGD) + (K × CoC) + ε

Here PD × LGD is the expected loss on the guaranteed exposure, K is the regulatory capital the guarantor must hold against it, CoC is the guarantor’s cost of that capital, and ε is administrative overhead. The task is to keep the premium low enough to preserve a positive spread for the pension investor, and the framework does so through three structural interventions.

First, the subordination of expected loss. The guarantee does not wrap the whole portfolio. Because the originating institution retains the 15 per cent first-loss position, that position absorbs the portfolio’s expected loss. The guarantee attaches only above the first-loss detachment point, so the guaranteed senior notes are exposed only to unexpected loss, the tail events that exceed the retained buffer. The expected-loss term facing the guarantor therefore falls close to zero and its required capital falls with it, which compresses the first two components of the premium.

Second, capital relief for the originator. The premium is borne by the originating institution rather than by the pension investor. The institution can carry it without a net loss because de-recognising the placed tranches frees expensive core equity capital that had been held against them, net of the charge that the retained position continues to attract under Section 5. Provided the guarantee premium is lower than the institution’s own cost of equity on the capital thus released, the institution is better off, and the underlying asset yield can be passed to the senior investor uncompressed. This is capital relief, and it should be described as such; the term arbitrage is best reserved for the regulatory concern it can raise, addressed in Section 5.

Third, concessional blending for specific vintages. Where a pool is heavily weighted toward climate transition or green infrastructure, global donor capital of the Green Climate Fund type can be used to subsidise the capital and administrative components of the premium. This widens the spread available to domestic investors and rewards the accumulation of green assets, aligning the framework with climate-finance objectives without distorting the credit assessment of the underlying loans. The contraction of concessional budgets documented in Section 1 sharpens rather than weakens this intervention: donor capital that has become too scarce to fund volume is precisely the capital that should be spent catalytically, subsidising the premium that unlocks a multiple of domestic institutional money rather than substituting for it.

One boundary of this section must be drawn precisely. The guarantee priced here attaches above the first-loss detachment point, and it is cheap for exactly that reason. The capital treatment of Section 5 requires a second and economically different protection, on part of the retained first loss itself. That protection cannot be priced on the logic above, because the first loss is where expected loss lives, and a rational guarantor prices any layer of it accordingly. The two guarantees are therefore priced separately, and the second is priced in Section 5, where the layer it covers is defined.

5. Capital mechanics: which regime binds, and what retention truly costs

The framework lives or dies on its capital treatment, and it is here that the argument must be most precise. Three questions must be answered together: which capital regime actually binds the institution, how the transaction frees capital under that regime, and what the retained first loss costs once its treatment is priced honestly.

Which capital regime binds

It is a category error to analyse every development finance institution as a Basel bank, because most are nothing of the kind. Three institutional types must be distinguished. The first is the minority of development finance institutions licensed as banks, which fall under national banking law and its Basel implementation; for these the Basel securitisation framework binds as a matter of law. The second, and by far the most common among African national and regional institutions, is the statutory development finance institution created by its own Act and standing outside the banking statute; the Development Bank of Southern Africa and the Industrial Development Corporation are examples of the form. No supervisor imposes Basel on these institutions. They shadow it voluntarily, and the reason is a market one: they fund themselves in domestic and international capital markets, and the bondholders and rating agencies that price their paper expect capital measurement and disclosure in a Basel-comparable language. Compliance is therefore voluntary in form and binding in effect. An institution that departed from its published capital policy would pay for the departure through spreads, ratings and market access rather than through supervisory sanction, which is a slower discipline but a real one. For these institutions the operative capital constraint is the rating agency’s own model, such as the risk-adjusted capital framework applied by S&P Global Ratings (2017), together with the board-approved capital policy the institution has committed to in its funding documentation. The third type is the multilateral development bank, which answers to no national supervisor at all and manages capital against rating-agency models and its own charter; this is the population the G20 review addressed (G20, 2022).

The leading precedent illustrates the distinction and is repositioned here accordingly. Room2Run is routinely cited as proof that securitisation frees development-bank capital, and it is, but the capital it freed was headroom under the rating agencies’ models, since the African Development Bank has no Basel supervisor; the transaction’s achievement was to transfer risk while preserving the triple-A rating (African Development Bank, 2018). That is precisely the regime most statutory institutions face. The analysis below is nonetheless conducted in Basel conventions, for two reasons: voluntary shadowing has made Basel the common measurement language across all three institutional types, and the rating-agency treatments, where they differ in magnitude, agree in direction, charging retained first-loss positions heavily and recognising protection from highly rated providers. An institution applying the framework should run the arithmetic under whichever lens its investors actually price, and Section 8 makes that determination the first task of implementation.

How capital is freed

When an institution holds a loan, that loan is a risk-weighted asset, and prudential rules, statutory or self-imposed, require core equity capital to be held against it. Core equity is the most expensive form of funding, because it is shareholder capital, often sovereign, that is locked against the exposure and cannot generate new returns. If a portfolio generates one billion dollars of risk-weighted assets against a 10 per cent capital requirement, one hundred million dollars of core equity is immobilised. Placing the senior and mezzanine tranches through a true sale removes the corresponding risk-weighted assets, so the capital held against them is no longer required and becomes available for new origination without recourse to a shareholder capital injection. What survives the transaction, and what the naive account ignores, is the capital consumed by the position the institution keeps.

The retained first loss and the two constraints

The first-loss position of 15 per cent stays with the originator deliberately, because an originator that sold its entire book would have little incentive to underwrite well, the classic moral hazard of originate-to-distribute. By keeping the first loss, the institution absorbs the earliest defaults, insulates the senior investor and the guarantor from expected losses, and signals that its origination can be trusted.

This is precisely where the capital treatment can defeat the transaction. Under the Basel securitisation framework, a retained first-loss position typically attracts a 1250 per cent risk weight or an equivalent dollar-for-dollar deduction from Common Equity Tier 1 capital (Basel Committee on Banking Supervision, 2016, 2019), and rating-agency capital models impose charges of similar severity on the same position. If the capital freed by placing the senior and mezzanine is clawed back by a full charge against the whole retained first loss, the net relief collapses toward a fraction of the headline and the rationale of the transaction weakens with it. The pension regulators pose the mirror-image constraint on the other side of the structure, and the two largest African pension markets illustrate its two forms. In South Africa, Regulation 28 under the Pension Funds Act caps a fund’s allocation to debt not listed on an exchange and to any single issuer, so an unlisted or unrated senior note competes for a narrow bucket; in Nigeria, the investment regulation administered by PenCom goes further, imposing minimum credit ratings on allowable corporate debt and requiring that infrastructure bonds carry credit enhancement in the form of guarantees from the federal government or eligible development finance institutions. The senior tranche clears these constraints because the guarantee lifts it to an eligible investment-grade rating, and in the Nigerian case the guarantee is a condition of eligibility rather than a comfort; a mezzanine sold without equivalent enhancement will struggle to find a home.

Pricing the resolution honestly

The mechanical escape from the punitive charge is well known: obtain unfunded credit protection on the retained position from a highly rated provider, so that under the substitution treatment within the securitisation framework the protected position carries the risk weight of the protection provider, between zero and 20 per cent for a top-band supranational, rather than the 1250 per cent of the position itself (Basel Committee on Banking Supervision, 2019). Applied carelessly, the escape extends to the whole first loss on the assumption that the protection is cheap. It cannot be. The first loss is the layer that carries the pool’s expected loss, and a rational guarantor prices protection on it at roughly that expected loss plus its own capital charge and overhead. Wrapping the entire layer therefore hands the guarantor approximately the yield the retention was earning, in exchange for a capital treatment, a trade that hollows out both the economics and the incentive the retention exists to preserve.

The resolution is to split the layer, and to price each part on its own risk. The junior sublayer is retained funded and unprotected. It is sized to the pool’s lifetime expected loss, subject to a retention floor whose source depends on the jurisdiction, and African jurisdictions approach the question differently from the European convention. The South African Securitisation Notice issued under the Banks Act (South Africa, 2008) disciplines the originator chiefly by limiting the credit enhancement and other support a bank originator may provide to a scheme, so as to preserve the separation on which de-recognition rests, rather than by mandating a minimum retention; other African markets sit at varying points between detailed notices of this type and near silence; and for statutory institutions standing outside the banking statute altogether, domestic securitisation rules written for banks may not apply at all. That gap strengthens rather than weakens the case for a self-imposed floor, because a retention no rule requires is the more credible signal. Where the domestic framework imposes no minimum, the structure therefore adopts 5 per cent of the securitised exposures as its default, borrowed from the European convention (European Union, 2017) for a practical reason rather than a legal one: international co-investors, guarantors and rating agencies treat that figure as the benchmark of aligned incentives, and a structure retaining less will be priced as though something is being concealed. This sublayer absorbs the expected loss, keeps the originator first in line for its own underwriting mistakes, and accepts the punitive capital treatment as the honest cost of skin in the game. The upper sublayer, the remaining 10 per cent of the pool in the default configuration, carries only unexpected loss above the junior attachment point. Protection on it therefore prices near the guarantor’s floor of capital and overhead, in the tens of basis points, and the substitution treatment applies to it legitimately, because the risk being substituted is tail risk, which is what highly rated guarantors exist to hold.

The arithmetic of the stylised pool then reads as follows. Before the transaction, the one billion dollar pool consumes one hundred million dollars of capital. After it, the placed 85 per cent consumes nothing; the funded junior sublayer of fifty million is charged in full; and the protected sublayer of one hundred million, at a substitution weight of 20 per cent and a 10 per cent requirement, consumes two million. Fifty-two million remains consumed, forty-eight million is freed, and the net release fraction of 0.48 that drives Section 2’s multiplier is exactly this number. The gap between the 85 per cent placed and the 48 per cent released is the price of genuine retention, and the framework reports it rather than hiding it. The sensitivity also runs in the right direction: an institution whose seasoned book carries lower expected loss can justify a thinner junior sublayer, subject to the retention floor, and earns a higher release fraction, so the capital arithmetic rewards exactly the underwriting discipline the retention is meant to protect.

Because the substitution treatment sits close to the boundary the term regulatory arbitrage patrols, the treatment should be settled in advance rather than discovered at issuance. For that reason the implementation plan in Section 8 mandates a capital relief audit that pre-clears the treatment with the national banking supervisor where one holds jurisdiction, and with the rating agencies whose models function as the effective supervisor where none does.

6. Why existing securitisation models are insufficient, and what is new here

Conventional securitisation was designed to improve bank liquidity, optimise regulatory capital, diversify funding and transfer credit risk to private investors. Its objective function is financial efficiency and the maximisation of shareholder value. A development finance institution operates under a different mandate, in which the objective is to maximise developmental additionality while preserving institutional sustainability. Applying a conventional structure without modification creates a predictable misalignment: once assets leave the balance sheet, the institution can be drawn toward maximising transaction volume rather than developmental outcome, and, absent an explicit reinvestment obligation, the liquidity released can drift toward lower-risk commercial activity rather than higher-impact projects. The institution slowly becomes a distributor of assets rather than a financier of development.

The Developmental Capital Velocity Framework addresses this by repositioning securitisation as a capital-recycling mechanism rather than a capital-exit strategy. The reinvestment covenant is the feature that redirects released capital back into additional developmental origination and prevents the drift that conventional structures permit. The contribution of the framework lies in redefining the objective that securitisation serves, from the maximisation of balance-sheet efficiency to the maximisation of developmental throughput, achieved while remaining compatible with the capital regime, statutory or voluntary, that the institution actually operates under.

The framework also advances on the leading real-world precedent. The African Development Bank’s Room2Run transaction of 2018 was the first synthetic securitisation between a multilateral development bank and private investors; it transferred the mezzanine credit risk on a one billion dollar non-sovereign portfolio to specialist private-credit investors, with the bank retaining the first loss and the senior layer and committing the freed capital to renewable energy (African Development Bank, 2018). Room2Run proved that risk transfer and rating-model capital relief were achievable for an institution outside any national supervisor’s jurisdiction. Developmental Capital Velocity differs in three respects that matter for scale and for domestic capital mobilisation. It distributes the senior tranche, the block whose de-recognition frees the most capital, rather than only the mezzanine. It places that tranche with domestic pension funds, mobilising local long-term savings rather than international private credit, which is why the partial credit guarantee on the senior notes is indispensable. And it binds the proceeds through a reinvestment covenant and reports the result through the velocity and multiplier measures computed on capital genuinely freed, converting a one-off balance-sheet event into a repeatable institutional discipline. In this sense the framework operationalises, at the level of a national or regional institution, the balance-sheet optimisation agenda that the G20 review of 2022 set out for multilateral banks (G20, 2022).

7. Velocity against prudence: the risks the framework introduces

A framework built on speed must confront the possibility that speed is itself a hazard. The history of originate-to-distribute finance before 2008 is a warning that when an originator expects to sell an exposure, its incentive to underwrite that exposure with care weakens, and that the faster capital turns, the more quickly weak underwriting can be scaled. An honest account of Developmental Capital Velocity must therefore treat prudence as a binding constraint on the objective rather than as an afterthought.

Three specific risks deserve naming. The first is underwriting erosion: the prospect of recycling can tempt an institution to relax credit standards in order to feed the pipeline. The retained junior sublayer counters this by keeping the originator first in line for the earliest defaults, and the capital arithmetic of Section 5 now reinforces the same discipline from the other side, because a deteriorating loss history forces a thicker junior retention, which collapses the net release fraction and with it the velocity the institution is trying to report. Under this design, bad underwriting is self-defeating on the institution’s own chosen metric.

The second risk is refinancing risk, and it deserves harder treatment than the usual nod to pro-cyclicality. The recycling engine assumes that each seasoned pool can be placed on schedule. If capital markets close, the institution holds assets it planned to sell, and if it has committed new origination against placements it expected to complete, it has recreated, in developmental form, the failure of the pre-2008 conduits and structured investment vehicles: long, illiquid assets funded on the assumption of continuous refinancing. A 20 per cent liquidity buffer absorbs slippage in issuance timing within a cycle; it cannot fund a multi-year greenfield pipeline through a closed market, and pretending otherwise would be the framework’s most dangerous self-deception. The binding rule is therefore sequencing, and it belongs in the covenant: new origination is committed only against capital already freed by completed transactions, and anticipated future securitisations are treated as options the institution may exercise, never as funding sources it may rely on. Under that rule a market closure caps velocity at the cycles already completed, an outcome that disappoints, and does no structural damage.

The third risk is additionality dilution: velocity measured only as financial turnover could be maximised by recycling capital into safe, low-impact assets, which is why the metric must be paired with the developmental-frontier covenant and with impact measurement per cycle, and why Section 2 insists the cumulative multiplier be read together with the mobilisation it embeds rather than as a claim about the institution’s own risk-bearing.

The framework’s guardrails are thus conditions of legitimacy rather than incidental features. The funded junior retention preserves the incentive to underwrite well and ties velocity to credit quality; the sequencing rule converts refinancing risk into bounded disappointment; the reinvestment covenant preserves additionality; the seasoning floor and liquidity buffer preserve resilience within the cycle; and the governance requirements of Section 9 preserve the whole against political interference. Velocity without these guardrails is balance-sheet optimisation wearing a developmental label, which is the outcome the framework is designed to prevent.

8. Implementation pathway and measurement

Implementation should proceed in five deliberate phases, each conservative enough to build the confidence of regulators, rating agencies and investors before the next is attempted.

Phase 1, capital-regime determination and feasibility. Establish, before anything else, which capital lens actually binds: the national Basel implementation for licensed-bank institutions, the rating agencies’ capital models and the institution’s own published capital policy for statutory institutions that shadow Basel voluntarily as a condition of market funding. Convene the banking supervisor where one holds jurisdiction, the rating agencies in every case, the pension regulator, legal advisers and prospective guarantors, and confirm that the split-retention structure of Section 5 qualifies for relief under the binding lens while satisfying investor eligibility. Confirm in the same phase the domestic securitisation requirements of the issuing jurisdiction, which across African markets range from detailed notices of the South African type to near silence, and settle the retention floor accordingly. Assess domestic pension-market capacity and complete preliminary due diligence on the vehicle. The capital relief audit belongs here.

Phase 2, pilot asset pool. Identify a small, diversified portfolio of seasoned assets with stable repayment histories and predictable cash flows, chosen to minimise model risk and to demonstrate proof of concept before any larger issuance. The pool’s loss history sets the thickness of the junior retention, and through it the net release fraction the pilot can honestly project.

Phase 3, structured issuance and placement. Issue the senior, mezzanine and split first-loss structure with both layers of credit enhancement finalised beforehand, supported by an investor-education programme for pension funds, asset managers and rating agencies.

Phase 4, performance monitoring and velocity assessment. Extend monitoring beyond credit performance to the measures that define the framework: the net capital release fraction achieved against the release projected, the speed of redeployment, the additional developmental projects financed, the mobilisation attributable to external investors, and the annual velocity and cumulative multiplier defined in Section 2. These provide the first practical measurement of Developmental Capital Velocity.

Phase 5, institutional scaling. Expand incrementally across asset classes, regions and sectors, allowing demonstrated performance rather than projection to earn the confidence of regulators and investors, and feeding the lessons of each issuance back into asset selection, retention sizing, guarantee pricing and reinvestment policy. The sequencing rule of Section 7 governs throughout: each expansion is funded by completed recycling, never by anticipated recycling.

9. Enabling conditions and failure modes

Like any institutional innovation, the framework depends on enabling conditions, and it should not be adopted where they cannot reasonably be met. It is unlikely to succeed where an institution lacks a sufficient volume of seasoned, performing assets to support predictable cash flows, because weak asset quality both undermines investor confidence and, through the retention sizing of Section 5, directly suppresses the release fraction the framework can deliver. It may prove impractical in jurisdictions with shallow domestic institutional capital markets, where pension funds, insurers and long-term managers lack the scale or the regulatory flexibility to absorb repeated senior issuance, though the direction of market infrastructure is favourable: the Johannesburg Stock Exchange's Forge 2031 strategy targets a pan-African digital marketplace by 2031 (Business Day, 2026), and exchange listing of the senior notes would work on both sides of this constraint, widening the allocation buckets available to pension investors, which favour listed instruments, and compressing the liquidity premium investors charge for paper they cannot trade. Announced strategy is not yet delivered infrastructure, and the framework does not depend on it; a functioning cross-border secondary market would, however, shorten cycles and diversify placement, converting a domestic market closure from an engine-stopping event into a rerouting problem. Uncertainty of capital treatment is a further constraint: where a supervisor, or for statutory institutions a rating agency, does not provide clear treatment of securitised exposures and split retentions, institutions may face charges that eliminate the benefit of the transaction, which is why early engagement is a prerequisite rather than an option. The framework also assumes the availability of credible enhancement providers whose guarantees are acceptable to investors and rating agencies, for the senior notes and for the upper first-loss sublayer alike; in their absence the senior tranche cannot reach investment grade and the retained position cannot escape its punitive charge, and demand and relief evaporate together. Finally, it requires strong governance in the originating institution, because failure to enforce the reinvestment covenant or the sequencing rule, deterioration in underwriting, or political interference in credit allocation would each convert the framework, over time, into the conventional balance-sheet optimisation exercise it was designed to transcend.

Stakeholder alignment

10. Conclusion

Development finance in Africa is constrained less by the absence of good projects than by the slow turnover of scarce public capital, and the constraint is tightening: with donor assistance in record contraction, grants shrinking faster than loans, and sovereign shareholders unable to recapitalise at need, capital that cannot be replenished must be recycled. Developmental Capital Velocity reframes the problem: it asks how many times a development finance institution can make the same capital work, gives that question two measures, a velocity that records the rate at which capital turns and a multiplier that records the cumulative origination the turning achieves, and computes both on the capital genuinely freed rather than on the assets nominally placed. The institution’s task is stated as it should be, a constrained optimisation in which velocity is the control variable and prudence supplies the constraints. The structure that raises velocity is built entirely from existing parts: a stratified securitisation, a guarantee that makes the senior tranche eligible for domestic pension capital, a split first-loss retention priced layer by layer, and a covenant that returns the proceeds to development. Its capital treatment is workable under each of the regimes an institution may actually face, whether Basel binds by law, by the voluntary discipline of market funding, or through the rating-agency models that govern where no supervisor does. Its principal dangers, that speed can erode prudence and that a recycling engine can seize when markets close, are met by the funded junior retention, the sequencing rule, the reinvestment covenant, the seasoning floor and disciplined governance. On honest arithmetic the reward is a doubling of developmental origination from the same capital, and more where underwriting quality earns it. The precedent of Room2Run and the direction of the G20 capital-adequacy agenda suggest that the mechanics are established. What the framework adds is the objective they should serve: a balance sheet that turns faster in the service of development, rather than one that merely grows.

References