Follow the Interest Rate From the Bond Market to Your Kitchen Table
If you have ever watched the Federal Reserve cut interest rates and then found that your mortgage quote went up that same week, you are not imagining things and nobody made a mistake.I wanted to work out exactly how a number decided in a bond market becomes the number on your loan paperwork. What are the steps, who sets each one, and how long does each leg take.
The answer turned out to be structurally different from what I expected. It is not one chain running from the Fed down to your kitchen table. It is two separate systems that never touch, and almost every piece of advice you read about timing a mortgage assumes there is only one.
Here is a core sample through each of them.
How a Treasury Yield Becomes Your Mortgage Payment
AND CREDIT RISK14.00
AND MBS SPREAD2.00
- There is no single interest rate that flows down to everything you borrow. There are two separate systems, and most people assume there is one.
- The Fed sets an overnight rate. Prime is that rate plus three points, by convention rather than judgement. Credit cards, home equity lines and most variable business credit sit on top of prime and reprice within about a billing cycle of any Fed decision.
- Mortgages are not on that circuit. They track the 10-year Treasury, which the bond market sets through buying and selling. The Fed influences it. The Fed does not set it.
- Which is why the policy rate can sit near 4 percent while the 10-year is above 5 and a 30-year mortgage is above 7. Those are not three rungs on one ladder.
- And the gap between the 10-year and the mortgage rate is not a fixed markup. It has averaged about 170 basis points since 1990 and has recently run closer to 200 or 230, depending on whose survey you read. That difference is real money and it has nothing to do with the Fed, the Treasury, or your credit score.
Two machines, not one chain
I went into this expecting to write about a chain. Treasury yield at one end, your kitchen table at the other, a few links in between. That is how it gets described and it is how I thought about it.
It is wrong, and the way it is wrong matters.
There are two machines. They start in different places, they are operated by different people, and they feed different loans. If you know which machine your debt is attached to, you can predict what it will do. If you do not, you will spend years waiting for a Fed decision to change a number that was never connected to the Fed in the first place.
Machine one: the Fed, prime, and everything that floats on it
The Federal Reserve sets a target range for the federal funds rate, which is what banks charge each other for overnight money. That is the only interest rate the Fed actually sets. Everything else in the economy is a reaction to it.
The first reaction is the most mechanical thing in American finance. The prime rate is the UPPER BOUND of the federal funds target range plus exactly three percentage points. Not approximately. Not usually. That convention has held since 1994, and the major banks move within a day or two of an FOMC decision. With the target range at 3.75 to 4.00 percent, the upper bound is 4.00, so prime sits at 7.00 percent. It moved there on September 17, 2026, the day after the Fed raised.
The number everyone quotes is the Wall Street Journal prime rate, which the paper defines as the base rate on corporate loans posted by at least 70 percent of the ten largest US banks. In practice that means the Journal reprints the formula once seven of the ten have moved.
Once you are on prime, the rest is markup. A home equity line is prime plus a margin set by your lender and your loan-to-value. A variable-rate business line is prime plus a margin set by your credit. And a credit card is prime plus a great deal more.
That 14 point gap between prime and the average credit card is worth staring at. It is larger than the entire rest of the column underneath it. When the Fed cuts a quarter point, your card rate falls a quarter point and the other fourteen points do not move at all.
Machine two: the bond market, and the loans that actually matter most
Now the other core sample, and the one almost nobody describes correctly.
A 30-year fixed mortgage is not priced off the Fed. It is priced off the 10-year Treasury note, and the Fed does not set that yield. Buyers and sellers do, in a market, every trading day, based on what they think inflation and growth and government borrowing are going to do over the next decade.
The 10-year is used as the benchmark rather than the 30-year bond for a practical reason: most 30-year mortgages are not held for 30 years. People move, refinance, or pay off early, so the effective life of a mortgage is far closer to ten years than thirty.
From there the chain is short. Your loan is bundled into a mortgage-backed security and sold to investors. Those investors compare the yield on that security against the yield on a Treasury of similar life. Whatever extra they demand becomes the spread, and the spread plus the 10-year is approximately your rate.
This is why you can watch the Fed cut rates and see mortgage rates go up the same week. It has happened repeatedly. The policy rate and the long end of the curve are different animals, and when the bond market decides that cutting now means more inflation later, long yields can rise on the news of a cut.
The layer that changes thickness
Look at the mortgage core again. The bottom layer is the Treasury yield, and everybody talks about it. The top layer is the spread, and almost nobody does, which is a shame because it is the part that has moved most.
Between 1990 and 2021 the spread between the 30-year fixed mortgage and the 10-year Treasury averaged around 170 basis points. Starting in 2022 it widened dramatically, exceeding 300 basis points in some weeks. It has come in since then but has not gone back to normal. Against a 10-year of 5.28 percent in early October 2026, the mortgage surveys were publishing anything from 7.28 to about 7.57 percent, which puts the spread somewhere between 200 and 229 basis points depending entirely on whose average you take.
Three things widened it and they are all still here.
The first is prepayment risk, and it is the dominant one. A 2026 Federal Reserve Bank of Boston paper, Why Mortgage Rates Exceed Treasury Yields, studied what it calls the coupon spread - the gap between the rates paid by mortgage-backed securities and Treasury yields - and found that the factors driving the prepayment option explain about 80 percent of its variation since 2006.
The logic is simple once stated. A Treasury pays you on a fixed schedule. A mortgage can be paid off whenever the borrower feels like it, usually at the worst possible moment for the investor, and without penalty. Buying a mortgage bond means selling somebody an option for free. The spread is what investors charge for having sold it.
The second is volatility itself. When the Treasury market swings hard from day to day, prepayment becomes harder to forecast and the spread mechanically widens. That is why mortgage rates can jump in a week when nothing at all happened to the Fed.
The shape of the yield curve matters here too, and the Boston Fed put a number on it: a one percentage point steepening of the curve, measured as the gap between the 10-year and the 2-year, reduces the spread by about 40 basis points. A flat or inverted curve signals that markets expect rates to fall, which makes the prepayment option more valuable and the spread wider. In early October 2026 that gap was only about 0.45 percentage points. Flat. Which is part of why the spread has stayed stubbornly wide.
The third is that the Federal Reserve stopped buying mortgage-backed securities. During quantitative easing it was an enormous, price-insensitive buyer that did not much care about prepayment risk. It has been shrinking its balance sheet instead, which means private investors have to absorb the supply, and private investors demand to be paid properly for the risk.
Put a number on it. Hold the 10-year still and move only the spread back to its historical 170, and on a 400,000 dollar loan the payment falls by about 81 dollars a month at the narrow end of the current range, or about 163 dollars at the wide end. Over thirty years that is somewhere between 29,000 and 59,000 dollars. At the 2023 peak of roughly 300 basis points the same arithmetic ran to about 358 dollars a month.
Nobody voted for any of that. It is not in a policy statement and no committee set it. It is the price of uncertainty, and you pay it monthly.
Why your quote does not move the day rates do
There is a lag between the bond market and the number your loan officer gives you, and it catches people out constantly.
This is also why the headline number you read and the number you are quoted rarely match. Different surveys sample different lenders on different days with different assumptions about credit score, down payment and points. A spread between survey averages of a quarter point or more at any given moment is completely normal and does not mean anyone is wrong.
Where everything else actually sits
Two machines covers most of it, but not all of it. A few loans sit awkwardly.
What to do with this
Three practical consequences, and they are the reason the distinction is worth knowing rather than merely interesting.
If you are carrying a credit card balance or a home equity line, your rate is attached to the Fed and will move with it, quickly, in both directions. Watching FOMC meetings is rational for you.
If you are waiting to buy a house, watching FOMC meetings is close to useless. Watch the 10-year, and watch the spread. A Fed cut with long yields rising is worse for you than no cut at all with long yields falling.
And if you are deciding whether to pay points to buy a rate down, understand that you are buying down a number that contains a risk premium which may normalise on its own. That is not an argument either way. It is an argument for knowing what you are actually purchasing.
The thing that genuinely surprised me here is how mechanical the first machine is and how little anyone says about it. Prime is fed funds plus three. That is it. There is no committee, no model, no negotiation - the banks just add three. Half the arguments I have read about whether the Fed controls consumer borrowing costs would evaporate if people knew that one fact.
The second surprise is the spread, and I think it is the most under-reported number in American personal finance. Everyone watches the 10-year. Almost nobody watches the gap, and the gap has been doing a lot of the work for four years now. Somewhere around half a point of the typical mortgage rate is compensation for uncertainty rather than for the cost of money. That is a policy outcome of a sort, just not one anybody announced.
What I would not do is treat any of this as a forecast. I am describing the plumbing, not predicting the water pressure. Plenty of people who understood the plumbing perfectly well have been wrong about rates for four years running.
Your credit card and your mortgage are priced by two unconnected systems. The card sits on prime, which is the Fed funds rate plus three points, and it moves within a billing cycle of any Fed decision. The mortgage sits on the 10-year Treasury plus a spread, and the Fed sets neither of those.
The spread is the part worth learning, because it is the part nobody quotes and it has been running well above its long-run norm since 2022. The difference between a historically normal spread and the recent one is roughly half a percentage point on every mortgage written in America - not because of policy, but because uncertainty costs money and somebody has to pay for it.
Filed under: General Knowledge
