Daniel sent us this one — he wants to understand where development finance institutions and development banks actually sit in the spectrum between pure philanthropy and commercial lending. And he's asking about concessional capital versus catalytic capital, two terms that get thrown around like they're synonyms when they're really not. The core question is: what does this institutional plumbing actually do, and why does it matter for anyone who cares about global development?
Two point five trillion dollars. That's the annual financing gap for the Sustainable Development Goals in developing countries, according to UNCTAD's twenty twenty-four figures. Philanthropy covers less than one percent of that. So the question isn't whether charity can fill the gap — it mathematically cannot. The question is what fills the rest, and the answer lives in this gray zone that most people have never heard of.
A gray zone where institutions lend money at rates that aren't quite charity and aren't quite capitalism. The financial equivalent of a mullet — business in the front, development mission in the back.
actually not a bad way to put it. Though I'd argue the mullet metaphor undersells the complexity. A mullet is two distinct things stitched together. Development finance is more like a gradient — you can be at any point along the spectrum, and the same institution often operates at multiple points simultaneously. The World Bank has its concessional arm, the International Development Association, which makes grants and near-zero-interest loans to the poorest countries. And it has its market-rate arm, the IBRD, which lends at commercial rates to middle-income countries. Same institution, same building in Washington, completely different financial logic depending on which window you're looking through. And as impact investing has gone mainstream — global assets under management hit roughly one point two trillion in twenty twenty-five — the plumbing behind it is becoming the bottleneck. Concessional capital, catalytic capital, blended finance structures — these are the mechanisms that make impact investing possible in places where commercial capital simply won't go on its own.
Let's map this territory. On one end, you've got pure philanthropy — grants, a hundred percent subsidy, no expectation of return. On the other end, pure commercial lending — market rates, full risk pricing, profit maximization. Where do these development finance institutions sit?
Somewhere in the messy middle. And to understand that middle, you need a spectrum. Philanthropy on the far left — grants, no repayment, hundred percent subsidy. Then concessional capital — loans or investments made at below-market rates, with an intentional subsidy built in. Then impact investing — market-rate or near-market returns, but with a double bottom line that includes social or environmental outcomes. Then pure commercial lending on the far right — maximize returns, manage risk, that's it.
The boundaries between those categories get blurry fast, don't they? If an impact investor is earning market-rate returns, how are they meaningfully different from a commercial lender who happens to be investing in a sector with good optics?
This is the definitional fight that has consumed entire conferences. The honest answer is that the boundary between impact investing and commercial lending is porous and contested. Some impact investors argue that intentionality is the differentiator — they're deliberately targeting social outcomes, even if the returns happen to be market-rate. The skeptic's response is: if the returns are market-rate, the intentionality is just marketing. You're a commercial investor who's good at PR.
Which is why I find the concessional space more intellectually honest. At least the subsidy is explicit. You know what you're paying for.
The institutions that operate in this middle space fall into three broad categories. First, Development Finance Institutions, or DFIs — these are typically bilateral, owned by individual governments. Think CDC Group from the UK, FMO from the Netherlands, Proparco from France. Second, Multilateral Development Banks, or MDBs — the World Bank, the African Development Bank, the Inter-American Development Bank. These are owned by multiple governments. Third, National Development Banks — like BNDES in Brazil or KfW in Germany. They're domestically focused but operate on similar principles.
The core tension that defines all of them?
They're capitalized by taxpayers but expected to operate with commercial discipline. They need to be profitable enough to be sustainable — most of them carry AAA or AA credit ratings — but not so profitable that they abandon their development mandate. It's a permanent balancing act. A DFI that loses money is politically unsustainable. A DFI that makes too much money gets accused of profiting off poverty.
The politics of that are fascinating. You've got a DFI reporting record profits, and the headline in the donor country's newspaper isn't "Development Bank Excels at Capital Management" — it's "Your Tax Dollars Are Making Bank Off the World's Poor.
And the DFI's response — which is technically correct but politically tone-deaf — is that those profits get recycled into new development lending. The institution isn't paying dividends to shareholders. It's building its capital base so it can lend more.
They're not just government-run charities with a banking license, which is what a lot of people assume.
That's one of the biggest misconceptions out there. These institutions maintain capital adequacy ratios, they manage loan portfolios with serious risk frameworks, and they access capital markets by issuing bonds. The World Bank's International Bank for Reconstruction and Development has been issuing bonds since nineteen forty-seven and has never missed a payment. That's not charity — that's institutional discipline. The IBRD bond is one of the most reliable instruments in global fixed income. Pension funds in Tokyo and Zurich hold World Bank bonds not because they care about development, but because they're AAA-rated and pay reliably.
Though the fact that they're disciplined doesn't mean they're not subsidized. And that subsidy mechanism — concessional capital — is where things get interesting. What actually makes a loan "concessional"?
It's simpler than people think. Concessional capital is about pricing. Imagine a solar company in Kenya that needs a loan to expand. The local commercial bank offers eighteen percent interest — that's what the market demands given the perceived risk. A DFI steps in and offers the same loan at four percent. That fourteen percentage point spread is the concession — a deliberate subsidy baked into the loan terms. The DFI is accepting a lower return than the market would require, not because they're bad at math, but because they're pursuing development outcomes that the market doesn't price.
That fourteen percent difference isn't just generosity. It's a policy tool.
The subsidy is the mechanism. You're not giving money away — you're lending it at terms that make a development project viable. The solar company in Kenya can now afford to expand into rural areas where the customer base is poorer and the infrastructure costs are higher. Without that concessional pricing, the project never pencils out.
Here's what I want to understand — how do they actually set that rate? Is there a formula, or is it basically negotiation?
There's a methodology, though it varies by institution. Most DFIs have a reference rate — typically based on their own cost of borrowing plus an administrative margin — and then they apply a subsidy element that reflects the development impact they expect. The OECD has a whole framework for calculating what's called the "concessionality level" of a loan, which is essentially the difference between the loan's face value and the present value of its repayments, discounted at a market reference rate. A loan with a concessionality level above thirty-five percent is generally considered concessional. Below that, it's just a slightly-cheaper-than-market loan.
There's actually a regulatory definition. It's not just vibes.
It's not just vibes. Though I should note, the OECD framework is primarily used for official development assistance reporting. In practice, DFIs have significant discretion. And the line between "concessional" and "near-market" is often more art than science, especially when you're lending in markets where there is no deep commercial benchmark rate to compare against. If there's no functioning commercial loan market in a given country — which is true in many of the places DFIs operate — what exactly is the "market rate" you're measuring your concession against?
Which brings us to the term everyone conflates with concessional: catalytic capital. Daniel's prompt specifically asks about the relationship between the two, and you can hear the frustration in the question — these terms get used interchangeably and they really shouldn't be.
I'm glad you flagged this, because the distinction matters enormously. Concessional capital is about pricing — below-market returns. Catalytic capital is about position in the capital stack — it sits at the bottom, absorbs the first losses, and by doing so, de-risks the investment for everyone above it.
Concessional is a price tag. Catalytic is a structural role.
They often overlap — catalytic capital is frequently provided on concessional terms — but they're conceptually distinct. I can have concessional capital that isn't catalytic, and catalytic capital that isn't concessional, though the latter is rare.
Give me an example of concessional capital that isn't catalytic.
A DFI makes a below-market-rate loan to a middle-income country for a road project that a commercial bank would have financed at a higher rate. The loan is concessional — it carries a subsidy — but it's not catalytic because it didn't unlock additional capital or de-risk anything for other investors. It just replaced commercial financing with subsidized financing. That's concessional without being catalytic, and it's exactly the kind of transaction that triggers the crowding-out concerns we'll get to.
Catalytic capital that isn't concessional?
Much rarer, but it exists. Imagine a foundation that provides a first-loss guarantee to a fund investing in affordable housing in Colombia. The foundation charges a market-rate guarantee fee — say, three percent of the guaranteed amount — so it's earning a commercial return on the risk it's taking. But the guarantee is still catalytic because it sits in the first-loss position and enables commercial investors to participate. The foundation isn't offering a subsidy on price; it's offering a risk structure that the market won't provide on its own. That's catalytic without being concessional.
Let's make this concrete with a full capital stack. Walk me through one.
Imagine a hundred million dollar fund that invests in water infrastructure projects in Senegal. At the top of the stack, you've got senior debt — eighty million dollars from commercial investors, pension funds, insurance companies. They expect market-rate returns and they want their money back first. Below that, you've got mezzanine debt or preferred equity — maybe fifteen million from impact investors willing to take a bit more risk. At the very bottom, you've got a ten million dollar catalytic tranche — provided by a foundation or a DFI — that absorbs the first ten percent of any losses.
If the fund loses five million dollars, the catalytic tranche eats all of it. The commercial investors at the top don't lose a cent.
That's exactly the mechanism. The catalytic capital functions as a cushion. It's first to take losses, last to get paid. By absorbing the risk that commercial investors won't touch — currency risk, political risk, liquidity risk, early-stage project risk — it changes the risk-return profile of the entire fund. The eighty million in commercial capital now faces dramatically lower risk, because losses have to eat through the entire ten million dollar cushion before they reach the senior tranche.
Catalytic capital isn't free money. It's expensive risk insurance, provided by mission-driven institutions that are willing to take losses commercial investors won't.
That's the right way to think about it. And this is where the magic of blended finance comes in. A ten million dollar catalytic grant from a foundation can unlock ninety million in commercial capital — a nine-to-one leverage ratio. That ten million didn't just fund a project. It changed the risk calculus for ninety million dollars that would otherwise have stayed in developed-market bonds.
Though the nine-to-one ratio is the optimistic case. What's the actual average?
Convergence, the blended finance network, tracks this. The average leverage ratio across blended finance vehicles globally is closer to four-to-one. Still significant, but a long way from nine-to-one. And that four-to-one includes a lot of vehicles that are mobilizing capital from other development finance institutions, not from purely commercial investors. When you strip out the DFI-on-DFI mobilization, the ratio of commercial capital to catalytic capital drops further.
The "billions to trillions" narrative — the idea that a little bit of catalytic capital can unlock a flood of commercial investment — is partly aspirational.
It's a vision, not a current reality. And the gap between the vision and the reality is where most of the interesting debates in development finance are happening right now.
Let's talk about a real example of this in action. The African Development Bank's Room2Run transaction.
This is a landmark deal from twenty eighteen, and it illustrates the mechanics beautifully. The AfDB had a portfolio of loans to African banks — loans that were performing fine, but were sitting on the AfDB's balance sheet and consuming its lending capacity. Institutional investors — pension funds, insurance companies — were interested in African exposure, but they considered the risk too high for direct investment.
The AfDB essentially said: we'll take the first losses, you take the rest.
It was a one billion dollar synthetic securitization — the first of its kind for a portfolio of African bank loans. The AfDB provided a first-loss guarantee, meaning it agreed to absorb the initial losses if any of the underlying loans defaulted. That guarantee transformed the risk profile of the portfolio. Suddenly, institutional investors — including Mariner Investment Group and the European Commission — were willing to buy in. The AfDB freed up a billion dollars in lending capacity, which it could then redeploy into new development projects.
The key word there is "synthetic." They didn't actually sell the loans. They sold the risk of the loans.
Right — a synthetic securitization transfers credit risk without transferring the underlying assets. The loans stayed on the AfDB's books, but the risk of default was partially transferred to private investors. It's a sophisticated financial engineering tool, and it worked. The transaction won multiple industry awards, and it's been replicated in other contexts since.
Here's the question I imagine a lot of listeners are asking: if these loans were performing fine, why did the AfDB need to offload the risk at all? Wasn't this just shuffling paper?
It's a fair question. The answer is regulatory capital constraints. Like any bank, the AfDB has to hold capital against its loan portfolio. Every loan on its books consumes a portion of its lending capacity. By transferring the credit risk to private investors, the AfDB reduced the capital it was required to hold against those loans, which freed up capacity to make new loans. It's not that the loans were bad — it's that they were tying up capital that could be deployed elsewhere.
It's balance sheet optimization, development-bank style.
And that's not a criticism — it's how banks work. The innovation was finding private investors willing to take African bank credit risk at a price that made sense for both sides. The catalytic element — the AfDB's first-loss guarantee — was what made that possible.
Concessional and catalytic capital can genuinely unlock private investment at scale. But here's the uncomfortable question: does it actually work in practice? Not in theory, not in the pitch decks — in the real world, with real projects and real outcomes?
The evidence is mixed, and this is where the conversation gets honest. Let's start with a success story. The Global Energy Transfer Feed-in Tariff program in Uganda — GET FiT — ran from twenty thirteen to twenty twenty. It used about fifty million dollars in concessional capital to subsidize feed-in tariffs for renewable energy projects. The result: one hundred seventy megawatts of new capacity, and critically, about seventy percent of those projects would not have been built without the subsidy, according to independent evaluation.
Additionality was actually demonstrated. They financed things that wouldn't have happened otherwise.
That's the gold standard for development finance — proving additionality. And GET FiT nailed it. The fifty million in concessional capital mobilized roughly four hundred fifty million in private investment. That's a nine-to-one leverage ratio, which is excellent by blended finance standards. The key design feature was that the subsidy was structured as a top-up payment per kilowatt-hour generated, not as an upfront grant. Developers only got the subsidy if they actually produced power. That aligned incentives beautifully.
It's not all GET FiTs out there.
Far from it. Let me give you the counter-example. Mozambique's tuna fishing fleet — a project funded in part by the European Investment Bank and other development finance institutions. Concessional loans were used to finance a fleet of fishing vessels that turned out to be commercially non-viable. The project failed. Mozambique was left with debt it couldn't service, the vessels sat idle, and the development impact was essentially zero — or negative, if you count the debt burden.
The white elephant problem. Concessional loans for large infrastructure projects that look great on a spreadsheet in Brussels and fall apart on the ground in Maputo.
This isn't an isolated case. The transaction costs of structuring blended finance vehicles can be prohibitive — often five to ten percent of the total deal size. When you're trying to mobilize a twenty million dollar fund, spending two million on lawyers and consultants to structure the capital stack starts to look questionable.
There's also the crowding-out critique. When DFIs lend at concessional rates, do they displace local commercial banks that can't compete with subsidized pricing?
There's evidence from India's renewable energy sector that suggests yes — domestic banks lost market share to concessional international lenders who could offer terms the local banks couldn't match. If a local bank would have made that loan anyway at eighteen percent, and the DFI steps in at four percent, you haven't added anything — you've just replaced a commercial lender with a subsidized one and distorted the local market in the process.
The DFI pats itself on the back for financing renewable energy, but the actual net impact is just a subsidy transfer from foreign taxpayers to the project developer, plus a weakened local banking sector.
That's the crowding-out critique in a nutshell. And it's hard to measure. How do you prove that a local bank would have made a loan that never happened because a DFI got there first? You're trying to observe a counterfactual.
Though I'd imagine there are cases where the local bank wasn't going to make the loan at any rate — where the DFI is filling a gap, not crowding anyone out.
And that's the additionality question. The challenge is that DFIs have an incentive to claim additionality even when it's dubious. Every DFI project document includes a section on additionality, and I have never read one that said "this project would have happened anyway without us." The incentives to overclaim are structural.
Which brings us to the measurement problem more broadly. How do you quantify "development impact" in a way that's comparable across projects?
This is the holy grail, and nobody's fully solved it. There are frameworks — the IRIS plus system from the Global Impact Investing Network, the Operating Principles for Impact Management — but they're still immature. A DFI might report that it "improved the lives of ten million people," but what does that actually mean? Measured against what baseline? Over what timeframe?
The incentives for DFI staff don't necessarily push toward rigorous measurement.
You've put your finger on something important. DFI staff face a career incentive to finance safe projects in middle-income countries rather than risky projects in the poorest countries. If you're a loan officer at a DFI, your performance review probably looks at your portfolio's default rate, not at whether you maximized development impact per dollar of subsidy. So you lend to a solar company in India — reasonably safe, decent optics — rather than to an agricultural cooperative in Chad where the development impact would be higher but the risk of default is also higher.
The safe middle path. It's the development finance equivalent of a cover band — you're not going to bomb, but you're not going to change the world either.
This creates what some analysts call the "concessional trap." Institutions that start with concessional lending often struggle to graduate to market-rate operations. Once you're known for subsidized loans, it's hard to pivot to commercial terms. Your staff, your systems, your institutional culture — everything is built around the concession.
Are there any examples of institutions that actually made that transition successfully?
The European Investment Bank is the rare success story. Between nineteen ninety-four and two thousand, the EIB shifted from primarily concessional lending to market-rate operations. It was a deliberate, multi-year transition that involved changing the institution's funding model, risk management systems, and staff incentives. Most DFIs never attempt it, and of those that do, most fail.
Why do they fail?
Because the skills required for concessional lending and commercial lending are different. Concessional lending is about understanding development impact and managing relationships with government counterparts and donor agencies. Commercial lending is about pricing risk accurately and competing with private banks. If you've spent your career doing the first, pivoting to the second is hard. And the institutional infrastructure — the IT systems, the risk models, the compliance frameworks — all need to be rebuilt.
What about the emerging innovations? I've been reading about outcome-based concessional capital, where the subsidy is contingent on measurable results.
This is the frontier, and it's exciting. Instead of providing a concessional loan upfront and hoping for the best, you tie the subsidy to verified outcomes. A pay-per-connection model for rural electrification, for example. The DFI provides concessional financing, but the subsidy only kicks in when a household is actually connected to the grid and verified by an independent auditor. If the connections don't happen, the concession doesn't materialize.
You're shifting the risk from the funder to the implementer. If you don't deliver, you don't get the subsidy.
And this aligns incentives in a way that traditional concessional lending doesn't. The implementer has a powerful reason to actually achieve the development outcome, not just to deploy the capital and move on. It's the bridge between the development finance world and the impact investing world — bringing the pay-for-performance logic of impact bonds into the concessional capital space.
The challenge, I assume, is that measuring outcomes is expensive and slow.
Nobody wants to be the DFI that publishes a report saying "we spent fifty million dollars and achieved nothing." But that's exactly the kind of transparency that outcome-based models require. You have to be willing to fail publicly.
Which brings us to the meta-question. Are we, as a global community, actually willing to be honest about the subsidies these institutions require and the tradeoffs they entail?
I think the answer is: not yet, but the conversation is shifting. For a long time, development finance operated in a kind of happy fog — everyone agreed the mission was noble, nobody looked too closely at the results. The rise of impact investing has changed that. When you're asking commercial investors to put money into blended finance vehicles, they demand rigor. They want to see the numbers, they want to understand the risks, and they're not satisfied with glossy annual reports full of smiling children.
The commercial investors are forcing a discipline that the development community never imposed on itself.
And there's a real tension there, because the development community's response is often: "you don't understand how hard it is to measure impact in fragile states." Which is true. But it's also true that "it's hard to measure" has been an excuse for "we're not really measuring" for decades.
For someone listening who's involved in this world — an impact investor, a philanthropist, a policymaker — what's the concrete takeaway?
Let me offer four. First, for impact investors: understand the capital stack of any fund you invest in. If there's no concessional or catalytic tranche, the fund is either taking on more risk than it's disclosing, or it's not reaching the hardest-to-finance projects. A fund that claims to be doing frontier-market impact investing with purely commercial capital is either being dishonest or is about to learn a painful lesson.
Second, for philanthropists: your grant dollars are most powerful when used as catalytic first-loss capital, not as direct project funding. A one million dollar catalytic grant can unlock ten million in commercial capital. A one million dollar project grant just funds one project. The leverage math is overwhelming.
Third, for policymakers: the next frontier is standardizing impact measurement so concessional capital can be priced rationally. The EU's Sustainable Finance Disclosure Regulation is a template, but it needs adaptation for development contexts. Right now, every DFI measures impact differently, which makes comparison impossible and allocation inefficient.
Fourth, the -insight: development finance is not a compromise between charity and capitalism. It's a distinct third category with its own logic, its own risks, and its own accountability structures. Treating it as charity-lite or capitalism-with-a-heart misses the point entirely. It's a system designed to do something neither charity nor capitalism can do alone — finance development outcomes in places where the market won't go and grants can't scale.
That system is under enormous strain right now. Climate adaptation needs alone are estimated at three hundred forty billion dollars per year by twenty thirty. The current blended finance market is around two hundred billion. That's orders of magnitude too small.
The question isn't whether these institutions work. It's whether they can scale fast enough, and whether we're willing to have an honest conversation about the subsidies they require.
There's an emerging phenomenon worth watching: concessional capital as a service. Specialized intermediaries — Convergence, the blended finance network, is the leading example — that match catalytic capital providers with commercial investors. They handle the structuring, the due diligence, the impact measurement. It's an attempt to professionalize what has historically been an ad-hoc, relationship-driven system.
Which could solve one of the biggest bottlenecks — the transaction costs. If you can standardize the structuring process, you can bring down that five to ten percent overhead and make blended finance viable for smaller deals.
The current system works reasonably well for hundred-million-dollar infrastructure projects. It works terribly for the five-million-dollar off-grid solar project in rural Uganda that might have enormous development impact but can't justify a million dollars in legal fees.
Where does this leave us? I keep coming back to the honesty question. We've built this elaborate institutional machinery — DFIs, MDBs, concessional windows, catalytic tranches, blended finance vehicles — and it does real things. It has built real markets. The off-grid solar industry in Africa, which now provides electricity to hundreds of millions of people, was substantially de-risked by concessional and catalytic capital in its early stages.
That's a genuine success story. Fifteen years ago, off-grid solar in sub-Saharan Africa was essentially nonexistent as a commercial market. DFIs and development banks provided the early-stage capital and the risk absorption that allowed companies like M-KOPA and d.light to prove their business models. Now it's a multi-billion-dollar industry attracting purely commercial investment.
Alongside that success, you've got the Mozambique tuna fleet. You've got the crowding-out problem in India. You've got DFI staff optimizing for career safety rather than development impact. You've got measurement frameworks that are still, charitably, works in progress.
You've got a financing gap that keeps growing. Two point five trillion annually for the SDGs. Three hundred forty billion for climate adaptation. The institutional machinery we've built is impressive, but it's not remotely adequate to the scale of the challenge.
Which is why the most important question isn't "do these institutions work." The most important question is "are we willing to be honest about what they actually do, what they cost, and what tradeoffs they require.
Because if we're not honest about the subsidies, we can't make rational decisions about how to allocate them. If we pretend concessional capital is just "smart investing," we'll under-subsidize the projects that need the most support. If we pretend it's just charity, we'll over-subsidize and crowd out commercial markets. The truth is in the uncomfortable middle.
That uncomfortable middle is exactly where Daniel's question placed us. These institutions aren't philanthropy, they aren't capitalism, and understanding them requires grappling with the tension rather than resolving it.
It's the financial equivalent of quantum superposition — they're both charitable and commercial until you observe them, at which point you have to pick a metric and live with the consequences of your choice.
That might be the nerdiest thing you've ever said on this show.
I regret nothing.
Now: Hilbert's daily fun fact.
Hilbert: In the seventeen twenties, a naturalist cataloging mineral specimens from the Seychelles archipelago documented a population of seventeen distinct meteoritic fragments embedded in granitic formations — meaning roughly one meteorite fragment for every twenty-seven square kilometers of the islands' exposed surface area at the time.
...right.
One meteorite fragment for every twenty-seven square kilometers. I'll be sure to bring a measuring tape next time I'm in the Indian Ocean.
The question that keeps me up at night — and I suspect it's the question that should keep the development finance world up at night — is whether we can scale this system honestly. Not whether we can scale it at all — the money will flow, the institutions will grow, the deals will get done. But whether we can scale it with genuine additionality, with rigorous measurement, with a willingness to admit when we're crowding out local markets or financing white elephants. That's the test.
The answer will depend on whether we treat development finance as a discipline or a religion. A discipline admits failure, learns, and improves. A religion defends its orthodoxies and punishes heretics. Right now, the development finance community contains elements of both.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop. If you want to hear more conversations like this one, find us at myweirdprompts dot com or wherever you get your podcasts.
Leave us a review if you enjoyed the episode — it helps other people discover the show. We'll be back next week with another prompt from Daniel.
