Evaluating the Mortgage Rate Lock-In Effect in the U.S. Housing Market: Evidence and Counterarguments
- Mar 17
- 9 min read
The mortgage rate lock-in effect, where homeowners with low fixed-rate mortgages are reluctant to sell due to higher prevailing rates, has been a subject of debate in recent housing market analyses. Some argue that lock-in is overstated or a myth, pointing to structural shifts like increasing housing tenure pre-COVID, the fact that about 40% of homes have no mortgage, and the role of first-time buyers not subject to lock-in. They also cite evidence of homeowners selling low-rate homes to rebuy at higher rates, as seen in the growth of the 6%+ mortgage cohort. These points correctly highlight that lock-in is not the sole driver of market dynamics and that sales have not collapsed entirely.
However, empirical evidence from multiple sources indicates that lock-in represents a meaningful incremental constraint on supply, suppressing inventory, reducing sales volumes, and contributing to elevated prices. This document reviews key arguments, incorporates data from academic papers, government reports, market surveys, and industry forecasts, and contrasts the U.S. experience with Denmark's model. All references are cited with footnote numbers at the end of the document.
1. Historical Sales Volumes and the "Suppressed" Nature of Recent Activity
Critics note that existing home sales reached 4.1 million in 2023, the lowest since 1995, but argue that if lock-in were dominant, volumes would have been even lower. They also correctly highlight that median housing tenure doubled pre-COVID, from 6–7 years in the 2000s to 10–11 years by 2019, per NAR data, suggesting structural stagnation rather than rate-driven effects.
While these pre-existing trends are acknowledged, research quantifies lock-in's additional impact. The FHFA's Working Paper 24-03 (revised 2024) uses National Mortgage Database data (1998–2024) to estimate an 18.1% drop in sale probability per percentage point rate gap, resulting in a 45% reduction in fixed-rate mortgage home sales in Q2 2024 and 1.72 million prevented sales from Q2 2022 to Q2 2024.¹
A 2024 Federal Reserve paper estimates lock-in accounted for 44% of the 2021–2022 mobility decline, equating to 719,000 fewer moves in 2022.² Without lock-in, sales could have approached 5.5–6 million annually. Bankrate's analysis notes 1.33 million prevented sales (mid-2022 to end-2023), with $511 monthly savings incentivizing homeowners to stay.³
Industry forecasters corroborate the persistence of this constraint. Redfin dubbed 2026 "The Great Housing Reset," projecting only 4.2 million existing home sales for the year, still near multi-decade lows, and noting that many would-be sellers continue to pull listings rather than accept lower offers. Zillow similarly finds a 17% inventory shortfall versus pre-pandemic levels entering 2026, despite gradual improvement.²⁸
The IMF's "The Pricing-Out Phenomenon in the U.S. Housing Market" (WP/23/1, 2023) models how lock-in reduces supply, pricing out lower-income buyers, with simulations showing a 1–2 percentage point rate increase reduces sales by 10–15% beyond baseline trends.⁴
Figure 1: Number of existing homes sold in the U.S. (2005–2024), with forecast to 2026. Source: NAR / Fannie Mae / Statista.
2. The Distribution of Mortgage Rates and Evidence of Movement
Observations show the 3–4% rate cohort declining from 40% to 30% of outstanding mortgages, with 6%+ growing, crediting repeat buyers for transacting despite low rates, indicating partial market fluidity, per ICE Mortgage Technology data cited in Bankrate.
Yet this does not fully counter lock-in's effects. FHFA data shows supply restrictions inflated prices by 7.0% cumulatively (Q2 2022–Q2 2024), offsetting a 5.6% demand dampening from higher rates.¹ Harvard's Joint Center for Housing Studies (JCHS) estimates a 1 percentage-point lower average rate boosts price growth by 8 points (2021–2023), explaining 40% of the price-rent divergence.⁵ The Fed paper finds lock-in raised prices 8% across Core-Based Statistical Areas and cut listings 10% in high-exposure areas.² Freddie Mac's November 2024 outlook quantifies average lock-in at $42,000 per loan (October 2024), with listings suppressed amid a 3.7 million unit shortage.⁶
Katz and Minton (2026) provide the most granular price estimate to date, finding that a one standard deviation increase in the rate lock incentive (equivalent to a 0.3 percentage point lower average outstanding rate) caused 2.6 percentage points higher nominal house price growth during the 2021–23 hikes. Their counterfactual shows that without lock-in, the price-to-rent ratio would have declined 12.3% versus only 4.4% with fixed-rate mortgages in place.²⁶
Summary: Estimated Price Impact of Lock-In Across Studies
The following table consolidates price impact estimates for quick reference:
Source | Estimated Price Impact | Time Period / Context |
FHFA Working Paper 24-03 | +7.0% cumulative price inflation | Q2 2022–Q2 2024 (supply restriction effect) |
Harvard JCHS (2023) | +8 percentage points of price growth per 1pp lower rate | 2021–2023; explains ~40% of price-rent divergence |
Federal Reserve Paper (2024) | +8% across CBSAs; listings cut 10% in high-exposure areas | 2021–2022 rate hike cycle |
Katz & Minton (2026) | +2.6pp nominal price growth per 0.3pp lower avg. rate | 2021–2023 hikes; counterfactual shows ARM market would have reduced P/R ratio by 12.3% vs. 4.4% with FRMs |
The MSCI Agency MBS Prepayment Model (2023) further discusses how slowed prepayments due to lock-in reduce MBS turnover and support higher prices.⁸
Figure 2: Share of outstanding U.S. mortgages below 3% and between 3.0–3.99%, 2013–2025. Source: FHFA / WolfStreet.
3. Composition of Real Estate Transactions: First-Time Buyers vs. Movers
A key argument is that first-time buyers, immune to lock-in, sustain the market. However, data shows their share has declined sharply, implying reduced mover activity amplifies lock-in's role.
NAR's 2025 Profile of Home Buyers and Sellers reports first-time buyers at a historic low of 21%, down from 34% in 2024 and a historical average of ~40%. Overall transactions remain low, with the first-time share contracting 50% since 2007 due to affordability and inventory constraints.⁹ Median first-time buyer age hit 40, up from 28 in 1992.¹⁰
This shift underscores lock-in's supply-side pressure, as fewer movers list homes, forcing more reliance on new construction or all-cash buyers (now ~30% of market, per NAR).
Figure 3: Percentage of first-time home buyers in the U.S., 1981–2024. Source: NAR 2024 Profile of Home Buyers and Sellers.
4. Household Surveys: Sensitivity to Rates as a Barrier to Mobility
Market surveys provide direct evidence of rate sensitivity. A February 2026 Storable survey of 1,000 U.S. adults found 46% of homeowners feel "trapped" by low rates, with 73% of mortgaged owners saying they'd consider moving if rates were transferable, and 31% would move immediately. This "golden handcuffs" effect delays life milestones like career changes and family formation.¹¹ ¹²
HomeServe's 2026 survey echoes this: 47% of homeowners stay put due to locked-in low rates, prioritizing this over high prices.¹³ NAR research confirms low mobility persists despite rate dips below 6% in early 2026, with uncertainty and job concerns compounding rate reluctance.¹⁴
Academic work reinforces the surveys: Katz and Minton (2026) compare owners buying just before and after rate spikes and find a 1 percentage point lower outstanding rate reduces overall moves by 42% and owner-to-renter transitions by 33%.²⁶ Fonseca and Liu (2024) estimate lock-in reduces household relocation by 1.4 percentage points annually, with a 0.4–0.5 percentage point drop in employment growth in affected areas.²⁷
The HousingWire/Altos Research analysis details how lock-in suppresses mobility, with estimates of 5–10% fewer moves annually.¹⁵
Figure 4: Household mobility rates by age group, 2019 vs. 2023. Source: American Community Survey / IPUMS.
5. Total Cost of Housing and Broader Market Dynamics
Rates are one factor amid rising prices, taxes, and insurance, and some low-rate sales occur, a fair acknowledgment of multifaceted costs.
Still, lock-in constrains supply. The Fed models 29% shorter time-on-market nationally in 2022, tightening conditions.² A September 2024 New York Times piece notes 60% of owners have rates below 4%, suppressing sales at 6.35%.¹⁶ By February 2026, rates fell to 5.98%, but January sales dropped, per NYT.¹⁷ Bankrate cites $3 trillion in cumulative savings keeping owners locked.³ Redfin notes that as of late 2025, sellers were pulling listings at the fastest pace in nearly a decade, further tightening already-constrained inventory.²⁸
"The Coming Rise in Residential Inflation" (Review of Finance, 2022) predicts 7% residential CPI/PCE inflation in 2022–2023 due to catch-up.¹⁸ "Comparing Past and Present Inflation" (NBER WP 30116, 2022) and "The Cost of Money is Part of the Cost of Living" (NBER WP 32163, 2024) argue mortgage costs distort sentiment and CPI, with lock-in amplifying effects.¹⁹ ²⁰
6. Contrast with the Danish Mortgage Model: What Non-Lock-In Looks Like
Denmark's system, similar in capital-market reliance, minimizes lock-in via match-funding: borrowers may repurchase their outstanding mortgage bonds at market value rather than par, monetizing the embedded capital gains when rates rise. A February 2026 paper by Berger, Jeong, Marx, Olesen, and Tourre uses comprehensive Danish administrative micro-data covering nearly 1.5 million homeowners per year (2010–2024) to demonstrate that household mobility in Denmark is largely insensitive to rate hikes.²¹
The contrast with the U.S. is quantitatively striking. Danish moving rates remained flat through the 2022–2023 rate surge, the estimated sensitivity of moving to a negative coupon gap in Denmark is economically negligible (0.12–0.20 bps per 100 bps of coupon gap). In the U.S., Fonseca and Liu (2024) estimate the equivalent coefficient is 57–120 bps, roughly an order of magnitude larger. To put this in concrete terms: the Katz and Minton (2026) finding that a 1 percentage point lower rate reduces U.S. moves by 42% would predict near-zero mobility impact under a Danish contract structure.²¹ ²⁶
The Danish data also reveal a behavioral pattern absent in the U.S.: refinancing hazards rise when rates fall (par prepayment) and again when rates rise sharply (discounted buy-back). In 2022–2023, approximately 100,000 Danish households executed discounted mortgage repurchases, realizing gains of up to 30% of face value. Danmarks Nationalbank (2023) separately documents that 20% of FRM borrowers refinanced in 2022's rate surge at 18% discounts.²² A 2018 comparative study highlights the structural stability this model provides.²³
Berger et al.'s equilibrium model shows that introducing a Danish-style repurchase-at-market option into U.S. mortgages would substantially reduce lock-in with equilibrium mortgage rates rising by only 2.6 basis points on average, a trivial cost relative to the mobility gains.²¹ A 2021 analysis of the Traditional Danish Mortgage Model details the underlying balance-principle architecture.²⁴
Figure 5: Danish FRM and ARM household moving rates, 2010–2023. Moving rates remain stable even as mortgage rates surged in 2022–2023, in stark contrast to U.S. lock-in patterns. Source: Berger et al. (2026).
Figure 6: Danish average coupon gap over time (top) and prepayment, refinancing, and moving rates as a function of the coupon gap (bottom). Note elevated refinancing at both positive (par prepayment) and deeply negative (discounted buy-back) coupon gaps. Source: Berger et al. (2026).
Figure 7: Danish household moving propensity as a function of the coupon gap. The flat line at negative coupon gaps demonstrates near-zero lock-in in Denmark, compared to a sensitivity 10x larger in the U.S. per Fonseca and Liu (2024). Source: Berger et al. (2026).
7. Policy and Forward-Looking Implications
Lock-in offsets monetary policy with asymmetric effects, per the Federal Reserve.² FHFA warns of persistence.¹ Denmark's evidence points to two primary reform pathways: (1) introducing a repurchase-at-market option into U.S. agency mortgages via FHFA modification of MBS guarantee terms, preserving existing TBA infrastructure; and (2) adjusting accounting and regulatory capital rules in the jumbo market to remove bank disincentives for accepting discounted repayments. The FHLB's analysis of in-substance defeasance for bonds provides a related precedent for liability-side market-value restructuring.²⁵
Berger et al. (2026) contrast the buy-back approach favorably against assumability and portability. Despite FHA and VA loans formally permitting assumptions since the 1980s, annual assumption rates have fallen below 0.05% of eligible mortgages, fewer than 3,825 VA assumptions and 2,221 FHA assumptions in 2023, against millions of eligible loans, underscoring that contractual availability alone is insufficient without removing lender discretion frictions. Buy-back rights, by contrast, are contractually self-executing at observable market prices and operate through existing refinancing infrastructure.²¹
Fonseca and Liu (2024) estimate lock-in's annual aggregate cost at 0.8–1.0% of GDP through reduced labor mobility. Redfin's 2026 outlook corroborates the macro weight of this constraint, projecting existing-home sales will reach only 4.2 million in 2026, still roughly 25% below the pre-pandemic 5.5 million pace, and describing the current era as a multi-year reset rather than a cyclical recovery.²⁷ ²⁸
Figure 8: Model-implied refinancing (left) and moving (right) hazard rates for U.S. FRMs, with and without a repurchase-at-market option. Introducing the buy-back option (dashed green) substantially raises moving hazards at negative coupon gaps, restoring mobility with only a 2.6 bps average increase in mortgage rates. Source: Berger et al. (2026).
In summary, while structural trends are foundational, lock-in has suppressed >1.7 million sales and inflated prices 7–8% per the academic evidence. Industry forecasters confirm the constraint persists into 2026. Denmark's institutional design, and the rigorous empirical evidence in Berger et al. (2026), demonstrates that lock-in is a product of contract design, not an inherent feature of fixed-rate mortgage markets, and that a targeted contractual reform could restore household mobility at minimal cost to borrowers or the mortgage system.
Footnotes
4) IMF Working Paper WP/23/1: "The Pricing-Out Phenomenon in the U.S. Housing Market"
8) Agency MBS Prepayment Model 2.1 (MSCI, 2023)
15) "What Everyone Needs to Know About Mortgage Rate Lock-In" (HousingWire/Altos Research)
18) "The Coming Rise in Residential Inflation" (Review of Finance, 2022)
19) "Comparing Past and Present Inflation" (NBER WP 30116, 2022)
20) "The Cost of Money is Part of the Cost of Living" (NBER WP 32163, 2024)
21) Berger, Jeong, Marx, Olesen, Tourre. "A Danish Fix for U.S. Mortgage Lock-in?" (February 2026)
22) Danmarks Nationalbank (2023): Refinancing Behaviour by Homeowners in Denmark
23) "Peas in a Pod? Comparing the U.S. and Danish Mortgage Finance Systems" (2018)
24) "The Traditional Danish Mortgage Model" (2021)
25) FHLB Memo (1992): Effect of In-substance Defeasance on FHLBanks
26) Katz, Justin and Minton, Robert. "Mortgage Rate Lock and House Prices" (January 2026)
27) Fonseca, Julia and Liu, Lu. "Mortgage Lock-In, Mobility, and Labor Reallocation" (NBER WP 32781, 2024)
Comments