Why Mortgage Rates Drift Away From Treasury Yields

Prepayment risk, hedging flows, and the negative convexity problem.

Treasury yields fall, and people expect mortgage rates to fall right behind them.

Sometimes that happens. Often it does not. A large move in Treasuries can show up in the news, yet mortgage quotes barely change. On other days, Treasury yields move only a little, but mortgage rates feel sticky or even move the wrong way.

That gap looks strange because the link seems simple. Treasuries are the basic interest-rate benchmark. Mortgages are loans built on top of that benchmark. So the path from one to the other should look clean.

But it often does not look clean at all.

The usual explanation is that lenders are slow to pass on lower rates. That story is easy to tell, but it stops too early. Mortgage rates are not set straight from Treasury yields. They pass through a market for mortgage bonds, and those bonds behave in a special way when rates move.

That is where the drift begins.

A Treasury bond is simple in one important sense. Its cash flows are fixed by contract. The owner knows when interest and principal are due.

A mortgage is different. A homeowner can often pay off the loan early. They can refinance, move, sell the house, or make extra payments. That means the investor who owns the mortgage may get principal back sooner than expected.

This is the hidden break in the simple story.

When rates fall, more borrowers can refinance. That makes old mortgages more likely to end early. When rates rise, refinancing slows, and the mortgage tends to last longer. So the life of the mortgage changes when rates change.

That matters because the thing being priced does not stay still while the benchmark moves. The benchmark moves, and the mortgage changes shape at the same time.

So the puzzle is not just that Treasury yields fall while mortgage rates lag. The deeper problem is that mortgages are not plain bonds. Their cash flows move around with the level of rates. That makes the pass-through messy even before any lender adds margin, fees, or other frictions.

The gap is not just about delay. It comes from the structure of the asset itself.

Most mortgages are not held one by one on a balance sheet for thirty years. They are pooled into mortgage-backed securities, usually called MBS. An MBS is a bond backed by many home loans. Investors buy that bond, and the mortgage rate offered to new borrowers depends in part on the yield those investors require.

The key issue is prepayment. Prepayment risk means the owner of the mortgage bond does not control when the principal comes back. The borrower does.

That changes the bond’s duration. Duration is a simple way to describe how much a bond’s price reacts to a move in interest rates. A longer-duration bond has more sensitivity. A shorter-duration bond has less.

In mortgage bonds, duration is not stable.

When rates fall, refinancing becomes easier and more attractive. Borrowers pay off old loans sooner. The principal comes back faster than expected. The mortgage bond gets shorter.

When rates rise, refinancing slows down. Borrowers keep their old loans longer. The principal stays out longer than expected. The mortgage bond gets longer.

This is the core of the negative convexity problem. Negative convexity means the bond changes in a way that is painful for the investor. When rates fall, the investor loses duration just when duration would have helped. When rates rise, the investor gains duration just when that added rate exposure hurts more.

That changing duration is not just a theoretical problem. It creates a risk-management problem for the people who own, finance, and intermediate mortgage bonds.

Banks, mortgage REITs, servicers, money managers, and dealers all tend to run limits on how much rate exposure they can carry. When the duration of their mortgage holdings changes, they often have to hedge that change somewhere else.

That somewhere else is usually the Treasury market or the interest-rate swap market.

This is where the drift between Treasury yields and mortgage rates starts to look mechanical instead of mysterious.

Start with falling Treasury yields.

At first, that should help mortgage rates. But lower yields also make refinancing more likely. As refinancing odds rise, mortgage bonds are expected to return principal faster. Their duration shrinks.

That creates a problem for anyone managing to a target level of rate exposure. Their mortgage holdings now have less duration than before. To get back to the target, they often need to add duration somewhere else. They may buy Treasuries. They may receive fixed in swaps. They may adjust hedge books in other ways that increase rate exposure.

Those trades are not a fresh opinion about the economy. They are a response to a change in the bond’s structure. The asset has shortened, so the hedge has to change.

Now turn the move around.

Suppose Treasury yields rise. Refinancing slows. Mortgage bonds are now expected to stay alive longer. Their duration extends.

That creates the opposite problem. Investors suddenly hold more rate exposure than they meant to hold. To cut that exposure, they may sell Treasuries, pay fixed in swaps, or reduce mortgage holdings.

So the mortgage system often has to trade in the same direction as the rate move. Falling yields can trigger hedge buying. Rising yields can trigger hedge selling.

That is why mortgage spreads can behave badly just when people expect a smooth pass-through. A spread is the extra yield investors demand over a benchmark such as a Treasury. If mortgage bonds become harder to hedge, or if their duration is moving around quickly, investors may demand a wider spread to own them.

That wider spread can absorb part of the Treasury rally. Treasury yields fall, but mortgage rates do not fall by the same amount because the spread stays wide or widens. The benchmark improved, but the mortgage bond became harder to manage.

This is one reason mortgage rates can drift away from Treasuries, especially in larger moves.

The effect can look even stronger when hedging happens in batches. Not every investor updates hedges every minute. Some rebalance after threshold moves. Some wait for model updates. Some act after the margin, mandate, or committee limits are hit. That means hedging pressure can arrive in waves.

When many firms adjust around the same time, the pass-through from Treasuries to mortgages can look weak, delayed, or even backward. No one has to coordinate. Similar assets under similar limits can produce similar trades on their own.

This mechanism does not explain every gap between Treasury yields and mortgage rates.

It matters less when refinancing is already blocked. If many borrowers cannot refinance because of credit limits, low home equity, closing costs, or paperwork frictions, then falling rates do not shorten mortgage bonds as much. The hedging need is smaller.

It also matters less when positions are light and hedge needs are small. If investors are not carrying much mortgage exposure, or if their balance sheets have room, the forced flows are weaker.

There are also times when mortgage rates are driven more by other frictions. Credit policy, underwriting standards, guarantee fees, servicing costs, and lender pipeline hedges can all shape the rate a borrower sees. In those cases, the Treasury market is only one piece of the final quote.

And this framework is narrow by design. It explains why the link between Treasury yields and mortgage rates can bend. It does not explain home prices, housing demand, or the whole credit cycle.

But within that boundary, the pattern is less puzzling than it first appears. Mortgage rates do not sit on top of Treasuries like a fixed markup. They sit on top of a bond that changes its length when rates move. Once that happens, hedging is no longer background activity. It becomes part of the transmission itself.