Macro shock: the AI displacement spiral
When intelligence becomes scalable, the macro transmission mechanism changes permanently. This is a structural read of that shift — not a set of recommendations.
The dominant narrative around artificial intelligence remains productivity-focused: AI automates tasks, margins improve, output rises. That framing is incomplete. A more rigorous macro read reveals a self-reinforcing loop with systemic implications.
The mechanism is straightforward. As agentic AI systems mature — capable of executing task chains, building software, and steering operational processes — firms substitute human knowledge work with automated alternatives. This is not speculative. By late 2025, capable developers with AI-assisted tools could replicate the core functionality of mid-tier software products in weeks. The “build versus buy” calculus has fundamentally shifted.
The macro risk emerges from what happens next. When firms cut headcount and redirect spending toward AI infrastructure, the savings do not recirculate through the consumption economy. Machines do not rent apartments, book vacations, or buy consumer goods. GDP may rise on paper — a phenomenon best described as “Ghost GDP” — while household purchasing power contracts.
Critically, this loop lacks a natural brake. In a classical business cycle, weaker demand eventually slows investment, prices adjust, and recovery follows. Here, AI investment accelerates precisely when the economy weakens, because it functions as cost substitution rather than incremental capital expenditure. A firm that once spent 100 million on salaries and 5 million on AI may shift to 70 million on salaries and 20 million on AI — total spend falls, but AI spend rises.
The transmission extends beyond software. Agentic systems strip friction out of intermediation-dependent business models — insurance renewals, travel booking, financial advisory, real estate brokerage — as algorithmic agents optimise relentlessly on price and efficiency. When friction disappears, the moats that sustained entire industries dissolve. And the scenario turns systemic if it reaches credit markets: private credit has expanded into software-adjacent deals underwritten on the assumption that recurring revenue stays recurring. AI-driven disruption challenges that assumption directly.
Inflation: CPI, PCE & the Truflation divergence
The gap between official US inflation metrics and real-time alternatives keeps generating debate. Understanding why they diverge is essential for judging whether monetary policy is calibrated correctly.
The three primary measures — CPI, PCE, and Truflation — broadly agree over long horizons. Over shorter windows, however, the divergence can be substantial, particularly during periods of rapid price change.
The decisive driver is methodology, not weighting. Truflation builds its housing component from mortgage rates, repair costs, property taxes, and insurance — the tangible costs a homeowner faces — and over-weights asking rents, which move faster than the average of all existing rents. Official statistics deliberately shifted away from that approach in 1983, recognising that it conflates investment cost with consumption, and instead use actual rental data from comparable properties.
The policy implication is significant: a simple Taylor Rule using Truflation data would have implied earlier, more aggressive rate hikes — potentially jeopardising the soft landing. For the current period, though, the divergence has narrowed, so the practical policy difference is limited. The takeaway: alternative measures are valuable as early-warning signals, but they do not replace the methodological stability of the official statistics that actually inform policy.
Structural trend: the US emigration shift
A structural trend is emerging that has not been seen at this scale since the Great Depression: net migration from the United States has turned negative. More Americans are leaving permanently than are arriving.
Official emigration data is incomplete, but international sources — residence permits, property purchases, university enrolments, work-visa applications — paint a consistent picture. Between four and nine million Americans now live abroad, and the trend is accelerating.
The drivers reach well beyond political polarisation. The primary force is economic: housing, healthcare, and education costs have risen to levels that make domestic life increasingly unaffordable for a growing segment of the middle class, while countries like Portugal, Mexico, Spain, and Thailand offer comparable or better quality of life at a fraction of the cost. Remote work has been the decisive accelerant — professionals earning dollar salaries discovered that geographic arbitrage sharply raises their purchasing power abroad.
The macro significance is clear. When a country’s middle class begins optimising globally rather than domestically, it signals that the implicit social contract — work hard, build wealth at home — is under structural strain, with downstream effects on consumption, tax revenue, housing demand, and the politics of redistribution.
Market implications & forward risk scenarios
The three themes are not isolated — they interact through feedback mechanisms that amplify macro risk.
AI-driven displacement weakens the consumption base. If inflation measurement understates the cost-of-living burden, policy may stay too restrictive for too long, accelerating the emigration of cost-sensitive households. Structural emigration, in turn, reduces domestic demand and tax revenue, narrowing the fiscal space available to respond to AI-driven labour disruption. Each theme quietly shrinks the room to deal with the others.
Framed as conditional scenarios, not predictions:
Accelerated displacement without policy response — agentic systems advance faster than institutions adapt; white-collar losses compound; private-credit portfolios face stress as recurring-revenue assumptions collapse.
Managed transition with friction-based deceleration — regulatory, legal, and organisational friction slows adoption enough for labour markets to adjust; productivity gains materialise gradually.
Inflation convergence triggers a policy pivot — official and alternative measures converge lower, giving the Fed room to ease; relief arrives, but perhaps too late if consumption damage is already embedded.
Geopolitical escalation compounds domestic stress — energy-supply disruption adds supply-side inflation, squeezing households already under strain and accelerating the emigration dynamic.
The limiting factor, across all four, is usually time — whether institutions can adapt before the feedback loops become self-sustaining.
We are not here to forecast which path arrives. We are here to make sure that when output rises while households feel poorer, you can see the mechanism connecting the two — and why the usual brakes may not apply.