How the engine works — the issuance channels

---

The one idea to hold onto

There is a single pot of new money each year, and it is tied to one thing: how much the real economy grew. Everything else — the channels, the modes, the dividend, the floor — is just how big you make that pot, and where you send it. Inflation isn’t one of the dials — it’s what the dials produce, falling out of where the money lands. (The one dial that raises it on purpose is the inflation-gap, KI — that’s Mode C.)

If you hold only that, the whole console reads cleanly: it's a control surface for sizing and routing one growth-tied budget, nothing more.

The base numbers (US calibration, all adjustable in the engine):

• M2 ≈ $22.4T — broad money.
• Real growth g ≈ 2% — how fast real output expands.
• Mᵀ ≈ $11.5T — the transactional slice of M2 (the share that actually moves goods prices). It's 51.35% of M2. This is the number that matters for inflation, and almost nobody separates it out — that separation is the whole trick.

Two derived anchors everything sits against:

• g × M2 ≈ $447B — the full growth budget (money growing in step with the whole stock).
• g × Mᵀ ≈ $230B — the price-stability locus: new money reaching the transactional circuit at exactly this rate holds goods prices flat. This is the leash.

---

The two pools (this is the part that's easy to miss)

Money created by the system lands in one of two circuits, and they behave completely differently:

• The transactional circuit (Mᵀ) — money people spend. Anything that lands here pushes on goods prices.
• The asset circuit (Mᵃ) — money locked into the citizen savings floor (equity, held). It buys assets, not groceries, so it does not push goods prices — except for a small, modelled leakage (spillover ≈ 20% of what's locked).

So the same dollar of new money is inflationary or not depending only on which pool it lands in. A dollar paid out as a spendable dividend hits Mᵀ. A dollar locked into the floor sits in Mᵃ and is nearly inert on prices. Hold that, and the modes stop looking like magic.

---

The four channels, and how each is calculated

Every year the engine computes these in order:

K1 — Citizenship. A one-time endowment into the floor for each new citizen.

K1 = k1  ×  (GDP / population)  ×  (new citizens that year)

k1 is a percentage of GDP-per-capita (default 2.5%). It's paid first, off the top of the budget. Set k1 = 0 and no citizenship endowment is created.

K2 — Growth rate. The throttle on the size of the whole growth budget.

Growth budget  =  k2  ×  M2  ×  g

k2 is the share of the growth-matched budget you actually issue. At k2 = 100% the budget is the full $447B; at k2 = 17.5% it's about $78B; at k2 = 0 there is no budget at all. This single dial sets how much new money exists. (K1 is then subtracted from it; whatever's left is the residual that gets split below.)

K3 — the dividend share (κ_d). This one is not a spigot — it's a router, and that's the most common point of confusion. It splits the residual budget between the two pools:

Dividend (to Mᵀ, spendable)  =  κ_d        ×  residual budget
Floor    (to Mᵃ, locked)     =  (1 − κ_d)  ×  residual budget

The two shares always sum to 100%. Moving the dividend up moves the floor down by the identical dollars — the total money issued does not change. So K3 is price-relevant (it decides how much lands in Mᵀ) but it is not a way to create more money. κ_d = 0 is all-floor; κ_d = 100% is all-dividend.

KI — the inflation gap. The only channel that issues money above the growth line.

KI  =  ki  ×  M2

This is additive — it's not carved out of the growth budget, it's piled on top, and it lands in Mᵀ. It is the only channel that deliberately creates inflation. Set ki = 0 and there is no above-line issuance. (Mode C uses it to fund a visible monthly dividend at the cost of ~2% inflation.)

---

Inflation is the output

Once the channels are set, the engine doesn't ask for an inflation rate — it computes one:

inflation  =  ( dividend  +  KI  +  floor spillover )  /  Mᵀ   −   g
                └──────── money reaching Mᵀ ────────┘

In words: take everything that actually reaches the transactional circuit, express it as a rate against Mᵀ, and subtract real growth. If money reaches Mᵀ at exactly the growth rate, the two cancel and prices are flat. Below it, deflation. Above it, inflation. The floor (K2/Mᵃ) barely appears, because locked money isn't chasing goods.

That's why the modes land where they do — and you can check each by hand:

Mode B and Mode D reach price stability by completely different routes — B locks 60% and pays 40%, D locks nothing and pays it all — yet both put about $230B into Mᵀ, which is the locus. That's the design working: the leash is on the transactional circuit, and there's more than one way to sit on it.

---

The point: every channel can be zero

Here is the thing worth showing anyone who hears "this is just printing money."

Set k1 = 0, k2 = 0, KI = 0. (K3 is then moot — there's no budget to route.) Walk it through the formulas above and every term is zero: no citizenship endowment, no growth budget, no floor, no dividend, no inflation-gap issuance. The system creates exactly zero new money. It is not a printing press with the dial stuck on "more." Zero is one of its settings.

And the consequence is the opposite of what the fear assumes. With no new money and a real economy still growing ~2%, the same money stretches over more goods, so the inflation formula gives:

inflation  =  0 / Mᵀ  −  g  =  −2%

A dollar gains about 2% a year. Zero issuance is the hard-money / fixed-supply corner — the gold-standard outcome — and the engine contains it as a limiting case. Mode A (mild deflation) is just a step in from there; Mode B/D (stability) a few steps further; Mode C (mild inflation) one step past stable. The framework isn't inflationary or deflationary by nature. It spans the whole range, and the dials decide — including all the way down to creating nothing at all.

That is what defuses the printing-money objection: the zero setting itself. And even at zero, the citizen comes out ahead of today, because it is today's dollar that quietly loses 2-3% a year — not this one.

---

Index versus dividend: ownership or income

The K3 router decides more than "locked or spendable." It decides whether a citizen's share of the new money becomes ownership or income — two different things with two different fates.

• The floor (the index side). The (1 − κ_d) share that stays locked isn't held as cash. It's used to buy equity in the domestic market index, and held. That makes each citizen a part-owner of the country's productive capital: it compounds with the market, it's wealth, and because it sits in the asset circuit it barely touches goods prices. This is the structural buyer — the system standing in the market every year, buying the index on citizens' behalf.

• The dividend (the income side). The κ_d share is paid out as spendable cash. It's income — consumed, hitting the transactional circuit, gone once spent. It builds no stake.

So κ_d isn't just a price dial. It's the choice between handing people a slice of the market and handing them a cheque. Mode B leans to ownership (60% floor); Mode A is almost pure ownership; Mode D is pure cheque — 100% dividend, no index bought at all (A* = 0).

How the floor buys the index — measured against total market cap

The index purchases are sized and tracked directly against the whole market. Three numbers do it:

A* — the structural-buyer flow. The dollars going into the index each year:

A*  =  K1 endowment  +  (1 − κ_d) × growth budget

This is the floor money — citizenship endowments plus the locked share of the budget.

c — the flow as a share of the market. A* measured against total index market cap:

c  =  A*  /  market cap

At the default $69T market, Mode B's A* ≈ $272B is c ≈ 0.39% of the entire market, per year — the annual bite the structural buyer takes.

ψ* — realized citizen ownership. Buying c% a year and holding for dur years doesn't simply stack to c × dur, because the market is also growing and citizens eventually draw their floors down. The realized share is a growth-discounted accumulation:

ψ*  =  c × annuity(g, dur),     annuity(g, dur) = (1 − (1+g)^−dur) / g

At the defaults (40-year hold, 2% growth), Mode B settles near ψ* ≈ 10% of the index owned by citizens collectively — leaving ~90% as tradable private float (1 − ψ*). The engine brackets this between two decumulation models and shows the band.

And the ceiling is the key line: no matter how long the hold, ψ* cannot run past c / g (here ≈ 20%). Ownership asymptotes — it does not compound without limit.

"If we didn't bound it" — why citizens don't end up owning everything

(The question below is the one a sharp reader asks first: what happens if the structural buyer is not bounded.)

This is the real tension in a permanent buyer, and it is worth being explicit about, because a sharp reader goes straight to it: if the system buys a slice of the index every year, forever, doesn't it eventually own the whole thing?

If nothing bounded it, yes. Naive accumulation — c × duration, with no growth-discounting and no draw-down — would climb past any level, and at the extreme a fully-captured floor would imply owning more than 100% of the market, which is impossible. That impossibility isn't waved away; it's the hinge of the feasibility argument (the structural-buyer paper's feasibility analysis, and the ownership-feasibility figure in the macroeconomic model, where naive full capture crosses 100% while the recalibrated design sits safely below).

Three things bound it, and all three are in the engine:

1. The c/g ceiling. Because the market grows at g while the buyer adds c, ownership asymptotes to c/g and stops — at the calibrated flow it never approaches 100%.
2. Decumulation. Floors aren't held forever; citizens draw them down in retirement, recycling shares back into the tradable float. That's why realized ψ* (≈10%) sits well below even the no-draw-down upper bound, c × dur.
3. Return attenuation. As citizens own more of the capital stock, the marginal return on it falls toward g. That feedback makes ever-larger ownership self-limiting instead of self-reinforcing.

The net: ψ* settles at a feasible share — meaningful ownership for citizens, with the large majority of the market still privately held and freely traded. The structural buyer is a bounded buyer by construction. If it weren't, the design wouldn't be feasible — which is exactly why the bound is load-bearing, not a footnote.

---

Everything above is exactly what the engine computes, line for line. The rigor behind each number — the Mᵀ separation, the 20% spillover, the price-stability locus, the realizable returns, the bounded ownership share — is set out in the architecture paper (the mechanical design), the macroeconomic model, and the structural-buyer paper. But the mechanics themselves are just this: one growth-tied budget; four dials that size it and route it between a spendable dividend and a bounded, index-buying floor; with inflation and ownership both read off the back.