Money Creation without Banks

A bank's core trick is writing two ledger entries at the same instant: an asset it now owns, your loan, and a liability it now owes, your deposit. Money was not moved, it was created. The bank typed it into existence. The Bank of England said this plainly in a 2014 paper, and central banks have said versions of it since, but it took me time to fully understand the implications. Money creation isn't a natural function of banking. It's a chartered privilege, backed by deposit insurance on one side and a central bank willing to lend against collateral on the other.

Now put that next to Africa's MSME credit gap. Banks holding a government bond do limited underwriting and get favorable capital treatment for it. Banks financing a trader do the opposite: real underwriting, real servicing cost, a loan officer who has to visit the shop and count the chairs. That loan officer costs nearly as much for a $150 loan as for a $15,000 one. Given a choice between an easy, well-treated asset and a hard, expensive one, the chartered creators of credit take the easy asset. This isn't a story about villainy. It's arithmetic, and arithmetic doesn't respond to appeals to purpose.

So the interesting question isn't how to make banks care more about MSMEs. It's whether credit creation has to be a bank's privilege at all, or whether it's a privilege that's currently bank-shaped for historical reasons, cost reasons, and nothing else.

The two things that gate creation

Two things stop a SACCO or a cooperative from doing what a bank does: the cost of underwriting fifty small loans instead of one large one, and the lack of a wholesale funding line that would let it lend more than its members' savings. Both are cost problems, not structural ones. AI collapses the first by reading transaction history, supplier ledgers, and stock turns at a cost that doesn't scale with loan size. Tokenized settlement collapses the second by letting a pension fund or a diaspora industrialist fund a facility directly, with covenant enforcement written into the disbursement mechanism instead of trusted to a relationship manager.

Put those together and the set of institutions capable of creating credit gets much wider. Not because the technology creates value out of nothing, but because the two costs that used to require a bank charter to absorb no longer require one.

Chartering the industrialist instead of the state

The natural instinct is to have governments fund this. Resist it. States running credit programs turn every sector choice into a political one, and the program inherits every incentive problem a subsidy has: it gets defended past its usefulness, it gets captured by whoever lobbies hardest, and it competes for exactly the domestic bank balance sheet that's already crowded out by sovereign borrowing. None of that is necessary here.

The better sponsor is a former Kenyan Boeing engineer who wants to fund an aviation parts factory in Nairobi and has a twenty-year view on why it pays off. He doesn't need Sereel or anyone else to decide that aviation parts matter. He decides that, and he puts capital behind it. The infrastructure's job is narrower: give him a legal wrapper that segregates his capital from every other sponsor's, and give him a disbursement mechanism that only releases funds against a real, verified purchase, a classroom lease for the plumbing guild, an invoice from the tooling supplier, not a lump of cash the borrower can spend on anything. Purchase-order finance, tokenized, with the verification automated instead of manual.

That structure keeps politics out for a reason, not by policy. Nobody at the infrastructure layer picked the sector. The sponsor did, with his own money and his own read on the economy he grew up in. A hundred sponsors making a hundred independent bets is a different animal from one government agency making one big bet, even if both are called "industrial policy" by someone looking for a headline.

Chartering by economic contribution

The Boeing engineer's problem isn't that he lacks capital. It's that no institution is chartered to evaluate whether his factory deserves it on the basis that actually matters: aviation parts substitution reduces Kenya's import bill, and a domestic supply chain for MRO work keeps foreign exchange in the country instead of sending it abroad. A commercial bank evaluates his loan the way it evaluates any loan, against collateral and cash flow history he doesn't have yet, because he hasn't built the factory. A development mandate would ask a different question: does this project's completion measurably change the country's trade balance, employment, or technical capacity. If the answer is yes, the charter should follow the project's own economic contribution. A borrower's existing balance sheet becomes secondary.

Central banks today guarantee the banking system: they backstop deposits, they lend against collateral in a crisis, they hold the payment rails together. Nowhere in that mandate does it say they guarantee economic development itself. This is where we need to fix things. A central bank willing to charter capital creation against a project's projected contribution to the real economy is doing something closer to what development finance institutions already attempt. It just skips the multilateral bureaucracy and the years-long approval cycle.

What doesn't get solved by any of this

Someone still has to bear the loss when the plumbing school's revenue doesn't cover the note. A smart contract can execute a disbursement and verify an invoice. It cannot absorb a default the way a bank's capital cushion does, and it cannot recapitalize itself after a bad quarter. That risk has to sit somewhere, and right now the honest answer is: with the sponsor, priced into the return he expects, not laundered away by the technology. Anyone pitching this without naming who eats the first loss is skipping the hard part.

The other thing that doesn't disappear is FX. A diaspora sponsor converting dollars into shillings to fund a local factory is touching capital-account policy whether he wants to or not, even if no government agency is the lender. "No politics" means no state as creditor. It doesn't mean no state involvement, full stop.

The mechanics of milestone verification are solvable engineering. The harder problem is finding one central bank willing to charter a pilot on these terms and document what happens. If you sit inside a central bank, sit on a monetary policy committee, or advise one on mandate reform, I want to talk about what that pilot looks like. Find me and let's build it.