When the Federal Reserve raised interest rates between 2022 and 2024, U.S. existing home sales fell by roughly 40% to levels last seen after the Great Recession. By textbook standards, prices should have followed sales downward. They didn’t. Real house prices stayed remarkably stable.
A new paper, “Unlocking Mortgage Lock-In: Equilibrium Effects in a Spatial Housing Ladder Model,” explains the puzzle. Co-authored by University of Illinois finance professor Julia Fonseca, Wharton finance professor Lu Liu, and INSEAD finance professor Pierre Mabille, the study argues that a single feature of the U.S. mortgage market — the inability to take a low fixed mortgage rate with you when you move — turns rising interest rates into an unexpected source of housing demand.
The paper builds on the authors’ earlier work documenting “mortgage lock-in”: the phenomenon where homeowners who locked in cheap mortgages during the historically low-rate years of 2020 and 2021 refuse to sell, because doing so would force them to give up those rates. In the new paper, Fonseca, Liu, and Mabille push the analysis from individual behavior to the housing market as a whole — and to the prices and rents that follow.
The Missing Downsizers
The natural first guess is that mortgage lock-in would mostly affect the volume of housing transactions, not prices. After all, a homeowner who decides not to sell is missing from the market twice: once as a seller of the current home, and once as a buyer of whatever home they would have moved into next. On the face of it, supply and demand both shrink by roughly the same amount.
The new paper shows that the cancellation is not symmetric. Using individual-level mortgage data, the authors find that lock-in disproportionately suppresses homeowners’ moves down the housing ladder — people in larger homes who would have sold and moved to something smaller. “The missing movers are disproportionately what we call missing downsizers,” Liu said. “They keep homes they would otherwise have given up, so the supply of existing homes shrinks by more than the demand for new ones.”
The result is a net increase in housing demand — enough, the authors estimate, to offset about a third of the price decline that higher rates would otherwise have produced.
The study’s findings are significant against the backdrop of declining sales of existing homes, Liu said. “The housing market is severely affected by mortgage lock-in.”
Why the Effect Shows Up in Existing Homes
A central insight of the paper is that the housing market in the headlines is almost entirely made up of transactions for existing homes, not new ones. Between 1981 and today, existing-home sales have accounted for 80% to 90% of all U.S. housing transactions; newly built homes make up the rest. That means the supply side of housing transactions are mostly affected by who decides to sell — and less so by who decides to build, at least over the shorter term.
Mortgage lock-in operates squarely on that margin. “Most of the public conversation about high interest rates focuses on builders and construction costs,” Liu said. “But existing homes are a much bigger part of the supply story, and lock-in directly affects whether those homes come onto the market at all.”
This is, the authors argue, an underappreciated channel of monetary policy. When the Federal Reserve raises rates to cool the economy, the usual story is that mortgages become more expensive, demand falls, and house prices follow. The new paper adds a counterweight: In an economy where most mortgages are 30-year fixed-rate loans that cannot be transferred to a new home, the same rate increase shrinks existing housing supply by locking would-be sellers in place.
“When central banks raise interest rates, they need to think about how mortgage market design interacts with such policies.”— Lu Liu
What the Model Says
To quantify these effects, the authors built a spatial housing ladder model — an economic framework that tracks households as they move between renting, starter homes, and trade-up homes, and between expensive and less-expensive areas. Households make these choices over their lifetimes, and the model produces equilibrium house prices and rents in each segment by matching supply and demand.
Calibrating the model to recent U.S. data, the authors simulate a temporary increase in mortgage rates from 3.5% to 6.5% — roughly the size of the 2022 to 2023 tightening cycle. In a world without lock-in, the rate hike pushes aggregate house prices down by about 16%. With lock-in, the decline is only about 11%. Lock-in, in other words, offsets about a third of the price decline that higher rates would otherwise cause.
Among mortgage borrowers, the model says mobility falls by about 25% in the period when rates rise — closely tracking the authors’ causal empirical estimate of 29%. Effects are largest in expensive metropolitan areas, where households tend to have bigger loan balances and are therefore more financially exposed to giving up a low fixed rate.
The effects on rents are also notable. Higher mortgage rates push some would-be buyers into the rental market, raising rents. In the model, lock-in modestly amplifies that increase by contracting rental supply in some segments. Taken together, monetary tightening is less effective at cooling house prices than a simpler model would suggest — and pushes up the cost of renting more than it pushes down the cost of owning.
Can Policy Help?
The paper evaluates one prominent policy response: a $10,000 tax credit to sellers of starter homes, similar to what the Biden administration proposed in 2024. The model finds that the credit does modestly raise mobility, but only a small share of the people who collect it move because of the policy. The vast majority would have moved anyway. The authors estimate the effective cost per induced move at roughly $650,000.
“That’s more than the price of a typical trade-up home in our model,” Liu said. “From an efficiency standpoint, demand-side housing subsidies are poorly targeted at the lock-in problem because most of the money goes to people who didn’t need it to move.”
The paper does not endorse a specific alternative, but Liu pointed to coordinating renegotiation and, going forward, mortgage market design as the more promising lever. Allowing borrowers to take their mortgage with them when they move — known as portability — or to assume an existing mortgage from a previous owner could in principle reduce lock-in directly, without the regressive incidence of a seller subsidy. The challenge right now is that existing mortgages do not come with these options, and so unlocking existing borrowers requires renegotiating existing contracts, which policy may be able to support.
“When central banks raise interest rates, they need to think about how mortgage market design interacts with such policies,” Liu said. “Otherwise, rate hikes in the U.S. will keep doing less to cool home prices than expected — and they push up rents, which feed directly into the inflation measures the Fed is trying to bring down.”
This article was partially generated by AI and edited (with additional writing) by Knowledge at Wharton staff. Read our AI policy here.