The Melt-Up Before the Break

The cleanest reading of the current macro picture is uncomfortable: the same conditions that make a near-term equity melt-up

plausible are the conditions that make the eventual reversal violent. US credit issuance is running roughly three times last year's pace, one-year real rates sit thirty basis points

Glossary · 25bp / bp

A basis point (bp) is one-hundredth of a percentage point (0.01%); 25bp equals 0.25 percentage points, the conventional increment of a single central-bank rate move. Policymakers quote rate changes, yields, and spreads in basis points to avoid the ambiguity of percentage-of-a-percentage phrasing.

Read more →

from negative territory, and a Middle East supply shock

Glossary · cost-push shocks

Cost-push shocks are inflationary impulses originating on the supply side — rising input costs, tariffs, wages, or energy prices — that lift the price level independent of demand. They worsen the output-inflation trade-off, forcing central banks to choose between accommodating higher prices or contracting activity.

Read more →

has begun to reprice sovereign curves. The Financial Policy Committee has flagged the geometry. The buy side is positioned for the upside leg of it.

Start with the mechanical case for higher equities. Capital Flows argues that the credit cycle, not sentiment, drives the cycle's inflection: falling real rates and tightening spreads push capital down the risk curve into high-yield, small caps and AI-levered tech. On their read, one-year real rates crossing into negative territory is the trigger, and the Fed's continued inaction is doing the work a cut would otherwise do. The positioning evidence — IGV (the US software ETF) implied-vol divergences, Russell range compression, accelerating high-yield issuance — is consistent with a market priced for the melt-up rather than against it.

The Bank of England's Financial Policy Committee sees the same setup and reads it as a vulnerability map. Valuations are particularly stretched in US AI-focused technology. Sovereign debt markets are characterised by concentrated leverage among a small number of hedge funds running similar strategies across jurisdictions. The Middle East conflict, in the FPC's framing, is a substantial negative supply shock that raises the probability multiple vulnerabilities crystallise simultaneously — energy prices and government bond yields have already moved in large, volatile steps. The melt-up thesis and the FPC's stress map are not contradictory. They are the same trade, viewed from opposite ends of the time horizon.

The credit cycle is economic statecraft, and letting it break is not a policy choice anyone can afford.

Where AI sits in the plumbing

The more interesting argument in the dossier is that AI is not simply an equity story layered on top of the credit cycle — it is part of the credit transmission itself. Capital Flows frames AI as a mechanism injecting credit into the economy without banks, by collapsing the upfront capital needed to start, market and build companies. That is a liquidity impulse expressed through a supply-demand shift in capital rather than through the discount rate. A heterogeneous-agent asset pricing result discussed by Marginal Revolution pushes the same direction from the other side: transformative AI can drive rapid GDP growth while pulling the risk-free rate toward zero and widening the equity premium

. If both are roughly right, long-end yields stop being a clean signal of where growth expectations sit, and the marginal AI-tech buyer is being subsidised twice — by cheap credit and by a structurally lower r* (the neutral real rate

Glossary · r* (natural rate of interest)

r* is the real short-term interest rate consistent with output at potential and stable inflation — the rate at which monetary policy neither stimulates nor restrains the economy. It is unobservable, must be estimated, and anchors how restrictive any given policy stance actually is.

Read more →

).

The dossier is one-sided on direction and the reader should treat it as such: every quantified forecast points the same way. Capital Flows assigns the one-year real-rates-go-negative-and-trigger-a-TINA-style-melt-up call medium conviction, and the continued-Fed-inaction variant slightly lower. The only genuine dissent comes from Crump and Gospodinov's rotated block bootstrap

— a term-structure recession model robust to changing economic and market structure — which puts US recession within a year of April 2025 at 0.75 in its modified form and 0.51 in its baseline. The reconciliation is uncomfortable but coherent: the melt-up and the recession call are not mutually exclusive, because the credit-driven equity leg can run precisely while the real economy is rolling over. That is the FPC's worry rendered as a trade. The base case we read from the dossier is that the marginal AI-equity buyer enjoys one more leg before the supply shock, the concentrated sovereign positioning, and the labour-market drag from AI productivity arrive at the same door.