A household that took a 30-year fixed mortgage in 2021 at 2.7% watched new mortgages go to 7% by 2023. Its house did not move on that news, and neither, in any way it could spend, did its loan. The payment stayed the same. A bond investor short a 2.7% coupon would have booked a large gain; the household cannot sell its mortgage, cannot close it at the market price, and keeps the gain only by staying put and paying 2.7% for years the market now charges 7%. So the loan has no price. It has three other properties, and each one drives a decision.
Leverage
The mortgage multiplies both sides of the house. A buyer with 20% down owns a house worth five times the equity; five years later, with amortization and some appreciation, the multiple is nearer two. The January 2006 buyer of a Los Angeles house at 20% down saw the equity wiped out in 2008 and earned 2.6% a year on the cash put in over twenty years. The January 2010 buyer of the same house with the same loan terms earned 13% a year on equity and was never underwater. The entry price and the next decade decided it, and the leverage decided how much.
Leverage does not change the liquid book; the house counts as a fifth of its value in stock exposure whatever the loan. It changes the reserve. The danger of a levered house is a forced sale in a bad year, so the household holds six to twelve months of spending and mortgage payments in bills outside the book, and tests them against a 30% fall in the house plus a year without income. The 2021 buyer’s $700,000 of equity survives that fall. A 2026 buyer with 20% down on the same house has $300,000 of equity and does not: a 30% fall leaves the loan $150,000 above the house, and only the reserve keeps the house until prices recover.
The rate, against today’s
The loan’s coupon is fixed. Whether it is a good liability depends on one comparison: the coupon against the ten-year Treasury yield, 4.7% in August 2026.
| loan | rate | against 4.7% | reading |
|---|---|---|---|
| 2021 buyer | 3.0% | below | keep the loan; a Treasury pays more than the loan costs, so a spare dollar goes to the book |
| 2026 buyer | 6.5% | above | prepay first; every dollar earns 6.5% risk-free, more than any Treasury, and lowers the leverage above |
Prepaying is buying back the loan at par. For the 2026 buyer it is also the best risk-free return available, and it reduces the leverage that makes the reserve test fail. For the 2021 buyer prepaying gives up a 3% liability to earn 3%, when Treasuries pay 4.7%. Keeping a cheap loan has one cost: moving hands it back, unless the loan is assumable, so the household that keeps it is also choosing to stay.
The refinance option
When rates fall, a borrower refinances into the lower coupon and cuts every future payment. The option is real and it belongs to the household, but it is neither free nor always available. It is paid for in the mortgage rate: 30-year mortgages have carried about 1.5 to 2 points over the ten-year Treasury for decades, and part of that spread is the lender’s charge for the borrower’s right to prepay. And it needs equity. Refinancing requires the loan at or under 80% of the house’s value, and rates fall in recessions, when house prices fall too. The 2006 buyer was locked out from 2008 to 2014 with the going rate a full point below the coupon, and refinanced three times afterwards. At a going rate of 6.95% neither sample loan is in the money today.
What the loan is not
It is not a bond position that needs a hedge, and it is not a source of income when rates rise. The household with a 3% loan and Treasuries at 4.7% earns the gap only if it holds Treasuries it could have used to prepay, and it is the Treasuries that earn it. The loan simply did not get more expensive. Counting that as a gain, or a short bond, invites the wrong instincts: hedging a position that cannot be closed, or holding long Treasuries against a liability that has no price.