Making Sense of the Markets: Why the Bond Curve (and Human Judgment) Still Matters

Earlier this year, around the time tariffs were originally announced on "Liberation Day," bonds were defying logic and leaving even seasoned traders scratching their heads. Today, things make a little more sense, but only just. Yields have fallen, MBS price discovery is less of a problem, and a clearer trading range has developed, yet investors are still wrestling to interpret a Federal Reserve that seems both cautious and conflicted.

At the heart of this story lies the term structure of interest rates, the relationship between time and yield on the Treasury curve. While it may sound technical, this simple measure is one of the most powerful tools for understanding where monetary policy, mortgage rates, and the broader economy are heading. I’ve long focused on the spread between the 2-year and 10-year Treasury yields. That spread, which has held near 50-basis points since April after being negative for more than two years, acts as a barometer of short-term monetary policy expectations and long-term risk sentiment. The 2-year note tends to reflect where markets believe the Fed funds rate will be a year from now, while the 10-year tells us how much uncertainty (and growth risk) is priced into the intermediate future. Even when rates are moving, as long as this spread stays relatively consistent, the market’s message is clear: the Fed isn’t likely to surprise.

But that doesn’t mean the Fed is in control. If anything, markets have been consistently ahead of FOMC voters. Since 2022, traders have repeatedly priced in an aggressive rate cut path before the Fed itself was ready to deliver. Each policy meeting has followed a familiar pattern: “buy the rumor, sell the news.” Fast money and algorithmic trading, what I call the "electronification of markets," has turned every Fed Day into a high-frequency guessing game. The result is volatility without clear direction or an obvious explanation, a market that oscillates between skepticism and overconfidence with mortgage rates usually suffering in the aftermath as the 2s/10s curve autocorrects.

Against this backdrop, Jerome Powell’s job has only grown harder. Inflation remains above the Fed’s 2 percent target and labor market indicators have either been clouded by statistical anomalies or delayed by the government shutdown, making it difficult to gauge the real state of the economy. Because of this, recent rate cuts deemed necessary to manage risk have generally felt more political than practical in the eyes of many market participants. The Fed’s cautious tone today reflects a lack of visibility though, not necessarily a lack of conviction. Policymakers simply don’t have enough reliable data to justify deeper cuts.

While Treasuries grab headlines, the more interesting story may be unfolding in mortgages. Since early September, pricing dislocations within the MBS coupon stack and aggressive competition among servicers have reshaped rate sheets. The race among lenders to retain existing borrowers (ed. note: "recapture wars") has intensified. Technology and AI are central to this fight. Having led a division focused on using AI in the mortgage space, I’ve seen firsthand how data and automation can reduce the cost of origination and transform loan servicing economics. The goal isn’t just to acquire borrowers, it’s to keep them.

This competition has compressed primary–secondary spreads and pushed mortgage rates lower than Treasury yields alone would suggest. But that dynamic cuts both ways: if Treasury yields rise and the 2s/10s curve spread meaningfully surpasses 50 basis points, mortgage rates would jump sharply. There’s now more room for rates to move up than down in the near term, especially when considering the extent of MBS outperformance vs. Treasury hedges since August.

So where do we go from here? I expect rates to stay range-bound; what I call a “bracketed” market. The 2-year Treasury is bracketed by 3.50 percent rally resistance and 3.75 percent sell-off support, and the 10-year Treasury by 4.00 percent resistance and 4.25 percent sell-off support. That implies maybe one or two more FOMC cuts over the next 12 months, but nothing more dramatic. Mortgage rates, currently around 6.375 percent, could settle near 6.125 percent by next fall even if spreads widen modestly. This assumes the market does not experience a major “flight to safety event”; sending mortgage rates into the 5-handle. That event would likely be a function of a drastically weaker employment situation and corresponding spike in credit delinquencies.

The bigger story isn’t about rates; it’s about structural change in the economy. The labor market is quietly weakening, with layoffs across major employers, from UPS and Amazon to Intel and Accenture. At the same time, automation and AI are reshaping how work is done, displacing lower- and mid-level coding and operations roles faster than many expected. The same technological forces transforming financial markets are also transforming the workforce that supports them.

In a world where algorithmic models move billions in seconds, it’s tempting to think human judgment no longer matters. But understanding the bond market still comes down to a few timeless principles: watch the curve, question the consensus, and never lose sight of the fundamentals. Whether you’re a trader, a lender, or a borrower, the smartest move is to stay grounded in data, but skeptical of hype. The market will continue to monetize fear and euphoria, but those who understand both the math and the mindset behind it will consistently outperform.

Adam Quinones is founder of dataQollab MBS Markets and former Global Head of Mortgages at Thomson Reuters/Refinitiv. dataQollab's new market monitoring dashboard offers institutional-grade, real-time TBA MBS data that's accurate, accessible, and affordable. Visit www.mbsmkt.com for more details.