Part IX - The Macro-Financial Channel#
Contents
This part exists because the rest of the document has a hole in it.
Part I prices the inputs. Part II predicts firm and labor behavior. Part III works through sectors. None of them answer what happens to interest rates, savings, asset prices, and public debt if the productivity claims are even partly true - and those variables feed straight back into the capex assumptions that everything else rests on.
The loop is real and this document had been treating it as open. Capital opened it with one observation - that success raises the cost of the capital required to continue - and nothing followed through.
Sections#
| Rates and returns | The neutral rate, the discount-rate channel, and why success is self-limiting | The feedback loop into Part I |
| Prices and the two economies | Deflation in cognition, inflation in everything physical, and what that does to policy | The defining political-economy fact of the 2030s |
| Assets and distribution | Who owns the inelastic complements, and what that does to wealth concentration | The distributional consequence of Game 3 |
| Fiscal | Tax bases built on labor income, meeting an economy that uses less of it | The slowest-moving and most consequential |
| Demography | The large, already-determined force pointing the other way | Revises Game 4's aggregate down and leaves its composition standing |
The one-paragraph version#
If AI delivers a productivity step-up, the real neutral rate rises - higher returns on capital everywhere pull the discount rate up with them. That makes the capex in Part I more expensive to finance exactly as it is working, which is self-limiting in a way no current projection prices. Meanwhile the price level splits in two: anything cognition-intensive deflates while energy, land, healthcare, and skilled trades inflate, leaving central banks with an index that averages two opposite signals and a policy tool that cannot address either. Value accrues to the inelastic complements, which are assets, and assets are held disproportionately by people who already hold assets - so the distributional consequence of consumer-surplus gains is wealth concentration alongside cheaper goods. And the whole thing lands on tax systems built to collect from labor income in an economy generating proportionally less of it.
Why this is the most under-analyzed part of the transition#
Three reasons, and each explains a different absence:
- It requires holding the technology story and the macro story simultaneously, and the two literatures barely overlap. Technology forecasters treat rates as exogenous; macroeconomists treat AI as a productivity parameter.
- The signs are ambiguous and the mechanisms fight each other. Higher productivity raises the neutral rate; higher savings from capital-income concentration lowers it; demographics lower it further. Net effect requires argument rather than assertion, so most analysis skips it.
- The timing is off-cycle. These effects land after the 2028–2032 diffusion window, which is past the horizon most forecasts commit to.
Which pages are conditional, and on what#
The five pages do not share an epistemic status, and reading them as if they did is the main way to misuse this part.
| Page | Conditional on | If the condition fails |
|---|---|---|
| Rates | The productivity step-up landing | Collapses to the demographic baseline - low rates, the 2010s continuing |
| Prices | Cognitive deflation being large | The split still occurs, an order of magnitude smaller and politically inert |
| Assets | Game 3's complement thesis | Concentration continues on its pre-AI trend |
| Fiscal | Only partially - demographics drive half of it | The squeeze arrives anyway, roughly a decade slower |
| Demography | Nothing | Unchanged - it is the unconditional floor |
The asymmetry is the point: the strains this part describes mostly happen anyway, on demographic power alone, while the upside scenarios need the technology to deliver. That makes Part IX less speculative than its conditional framing suggests - it forecasts strain with an AI multiplier attached, not strain that AI invented.
The three feedback edges#
What makes this part structural rather than appended: three of its conclusions feed back into assumptions made earlier in the corpus. Rates reprices the capex in Capital - the financial governor. Prices supplies the political fuel that C5 tracks, which is the channel through which macro discomfort becomes regulation years before any capability rule does. And Fiscal sets the budget inside which every state response in State capacity has to operate. A reader who skips Part IX will find the rest of the document quietly assuming things this part is responsible for defending.
What this part is not#
It is not a market forecast, and nothing here should be read as one - but note that one page in it, demography, is the opposite of conditional. It is the most certain thing in the entire corpus, and it argues that the labor conclusions elsewhere are too pessimistic in aggregate while leaving the compositional ones intact.
The claims are about directions and mechanisms under a stated productivity assumption, and the productivity assumption itself carries only ~20% by 2030 in Part V row 2 (re-scored down from 25% in round 8 - measured TFP is a high bar once quality-adjustment gaps are priced).
The conditional structure matters: most of this part is what happens if the optimistic case is right. That is worth working out precisely because the rest of the document spends its length on constraints, and the constrained case has no interesting macro consequences at all - it just looks like the 2010s continuing.
U7 is the adversarial twin of this part#
Complement erosion attacks the distributional pages (assets, prices) without killing the productivity story. If the scarce list has a short half-life, Part IX's concentration claims overstate late-2030s inequality even while early-window B7 still fires. Date-stamp which complement row is doing the work; do not treat the part as a permanent political-economy map.