Bond Traders Aren't Flying Blind on the Fed - They're Flying Into Tight Liquidity

Everyone's writing about Kevin Warsh's silence and pretending that's the only thing keeping bond traders up at night. The headline story from the July 28-29 FOMC meeting is that the new Fed chair - only in his second meeting - refused to provide any forward guidance on rates. "I understand the desire for rolling forecasts," Warsh told reporters, "but for our part, we need to observe market reaction to developments direct and unfiltered."

Bond yields exploded. The 30-year Treasury topped 5.2%, its highest level since 2007. The 10-year jumped more than 7 basis points to 4.677%.

The narrative is that traders can't price the Fed anymore because Warsh has removed the signal. And sure, that's part of it. The committee voted 9-3 to hold rates at 3.50%-3.75%, with three members wanting a hike. The June dot plot showed only one policymaker anticipating lower rates by year-end. Nobody knows what's coming next.

But Warsh's communication policy didn't break the plumbing. The plumbing was already tightening, and most commentary hasn't even looked under the hood.

Here's what matters more than whether the Fed speaks or stays quiet. The cash buffer that absorbed Treasury issuance in 2023 has been largely drained. When the government borrowed then, the money came from the overnight reverse repo facility - that giant pool of idle cash parked at the Fed by money funds and institutional investors. Today, that pool is mostly gone. The same deficit, the same Treasury bills, different funding source. The borrowing now draws from bank reserves instead of idle cash. Same deficit. Same bills. Different impact.

Repo market data from the first half of 2026 already confirmed the regime shift. The Fed's assumed ceiling on Treasury repo rates has been broken on a regular basis since 2025. The cash buffer that absorbed Treasury issuance in 2023 has been largely drained. Dealers have reported "challenging" stressed market periods - not September 2019 levels, but real enough to make them question what comes next.

And this is where the credit market tells the story the equity headline hides. Investment-grade spreads were sitting at just 74 basis points over Treasuries at the end of the second quarter - the tightest level in the first percentile over a 20-year lookback. That means investors were demanding almost no premium over Treasuries to lend to corporations. Fully compressed. Then came July.

LQD, the investment-grade corporate bond ETF, is down 2.2% over the last 20 days and 3.6% year-to-date, trading near its 52-week low of 106. HYG, the high-yield ETF, is down 0.3% over 20 days and 1.4% YTD. Meanwhile, SPY is up 9.5% year-to-date and 1.1% over the last five days.

That divergence is the dispersion signal. Equities are marching higher on compressed volatility and options-driven flows while credit is repricing risk upward. And if you look at the options structure on SPY, the put-to-call open interest ratio sits at 1.95 - nearly two puts for every one call. Traders are hedging while buying. They don't trust the strength either.

Now here's the mechanism that connects the plumbing to the bond selloff, and it's the part most people miss. When reserves tighten, funding costs rise. When funding costs rise, dealers who finance inventory in repo face higher costs. When dealer financing gets expensive, they reduce their market-making capacity. When market-making capacity shrinks, bid-ask spreads widen and volatility amplifies. Which means a Treasury selloff that starts on Fed uncertainty gets mechanically amplified by a dealer book that's already stretched thin on liquidity.

Warsh's silence didn't cause the selloff. It just removed the cushion that would have normally softened the move. If the Fed had signaled a September pause, bond traders would have had a floor to price against. Instead they got nothing - and with nothing to anchor them, the liquidity squeeze became visible all at once.

The closest historical parallel isn't 2013's taper tantrum or 2018's rate-hike volatility. It's September 2019. Reserve balances had been drifting lower for months. Treasury settlements and tax payments hit simultaneously. Overnight repo rates spiked well above the Fed's target range. The tape was quiet in the days leading up to it. The plumbing broke first, rates followed, and equities absorbed the shock within hours of the Fed's emergency response.

The difference today is that Warsh has explicitly declined to give the market those early signals. He won't telegraph the response. So traders are left reading repo rates, TGA flows, and reserve balances without a roadmap.

Yes, we could still go higher in equities. The bubble case at 25 times forward earnings puts the S&P 500 somewhere above current levels. But do you really want to play for upside when credit is widening, the put-to-call ratio on SPY is approaching 2-to-1, and the plumbing has been draining for two years?

Here's the conditional chain for the next move. If Treasury issuance continues to draw down reserves without offsetting RRP inflows, dealer funding costs stay elevated, and credit spreads keep widening as a mechanical function of tighter liquidity. If the Fed responds by cutting the standing repo facility rate or increasing repo operations, the plumbing loosens and spreads can stabilize. If Warsh pivots and provides forward guidance - a September hold or cut - the bond market gets its anchor back, and the selloff has a chance to reverse.

The bond market isn't flying blind. It's reading the plumbing and finding it thin. The question for the rest of us is whether we're going to keep watching the equity headline while the funding markets tell a different story.

What to watch: SOFR versus the Fed's interest-on-reserves rate - if that spread widens, funding is getting tighter. Treasury General Account balance - a rising TGA drains reserves. And the IG OAS from here - if 74 basis points turns back toward 90 or 100, the credit repricing is accelerating, not reversing.

These views are my own and do not constitute investment advice.