Chicago’s house sales fell 22 percent from 2018–19 to 2023–24, and they fell furthest in the community areas where more of the homes are owned by people 35 to 54: down 29 percent in the 15 areas with the largest share of those owners, against 11 percent in the 15 with the smallest, or 18 percent once sales by companies are set aside. Areas where more of those owners have no mortgage, and no cheap loan to protect, weren’t spared: at similar price levels they lost more sales too. And once age and price are counted, neither the share of owners with a mortgage nor the 2020–21 refinancing wave goes with a deeper drop. Our theory: the lock-in that matters is the loan a move would take, not the one an owner already has.
Cook County Assessor sales file, read September 30, 2026 · federal mortgage records (HMDA), 2018–2025 · Census American Community Survey · the comparisons and the theory are oursWhat is the lock-in effect in real estate?
It’s the pull that keeps an owner with a cheap mortgage from moving. On November 16, 2026, realtor.com’s research team put it this way: selling means giving up cheap financing for a more expensive loan on the next home, while an owner with no mortgage can sell on their own schedule. Its analysis of the 2025 American Community Survey expects places where owners rely more heavily on mortgages to be more sensitive to rate changes, and places with more outright owners to feel smaller effects. It puts 63.2 percent of homeowners in the Chicago metro area in homes with a mortgage, against 59 percent nationwide; in the city the 2025 survey puts it at 63.7 percent.
The cheap loans are real. Freddie Mac’s weekly survey of 30-year fixed rates averaged 2.96 percent in 2021, 5.34 in 2022 and 6.81 in 2023; it has averaged 6.54 so far in 2026 and stood at 7.11 on November 12. In Chicago, lenders made 91,630 refinance loans in 2020 and 2021 on homes their owners live in, by the federal mortgage records, at a median rate of 2.99 percent. If those loans are what kept Chicago owners from selling, the areas where the most owners carry a mortgage, and the most refinanced, ought to show the deepest drops, and owners with no mortgage ought to have kept selling.
Where Chicago house sales fell the most
The Cook County Assessor’s sales file records 26,612 house sales in Chicago in 2018 and 2019 that pass its screens for ordinary, arm’s-length sales, and 20,849 in 2023 and 2024. Houses here means single-family homes; condominiums and two-to-six-flats are left out. We took the 62 community areas with at least 100 house sales in 2018–19 and set each one’s change beside who owns the homes there, from the Census Bureau’s 2020–2024 survey. The survey sorts owner households by the age of the householder, the person in whose name the home is owned; “owners 35 to 54” on this page means owner households whose householder is that age, and each such share is of all the area’s owner households.
Of the measures on this page, age sorted the areas most cleanly. Across the 62, the correlation between the share of owners 35 to 54 and the change in sales is −0.61, on a scale where −1 would be a perfect downward line and 0 no relationship at all. The table also sorts the areas by their share of owners 65 and older, and its last column sets aside sales by companies, which account for much of the older areas’ smaller drop.
| The 15 community areas with | Share of owners in that group | House sales 2018–19 | House sales 2023–24 | Change | Change, sellers other than companies |
|---|---|---|---|---|---|
| the most owners 35 to 54 | 42 to 55% | 5,087 | 3,593 | −29.4% | −29.1% |
| the fewest owners 35 to 54 | 24 to 32% | 6,219 | 5,550 | −10.8% | −18.2% |
| the most owners 65 and older | 38 to 55% | 6,711 | 5,792 | −13.7% | −22.5% |
| the fewest owners 65 and older | 12 to 24% | 5,262 | 3,888 | −26.1% | −25.9% |
The raw gap opened in the same years rates rose, though the timing alone doesn’t show that rates caused it. Measured against each group’s 2018–19 average, the areas with the most owners 35 to 54 sold 94 percent as many houses in 2020 and the areas with the fewest 96 percent; in 2021, 111 and 109. In 2022, as the average rate climbed to 5.34, the two read 95 and 103; in 2023, 70 against 91, and in 2024, 71 against 88. The 2025 sales the county has recorded so far keep the same order, 71 against 92.
Did the areas with the most mortgages lose the most sales?
Taken alone, a little: the correlation with the share of owners carrying a mortgage is −0.32. But that share moves with price and with age. Cheaper areas have fewer mortgaged owners (+0.62 between the two), more owners past 65 and smaller drops. The fair question is whether mortgages go with a deeper drop once age and price are both counted, and they don’t: the correlation that’s left is +0.10 or −0.02, depending on how price is adjusted. Refinance loans per owner household, which track price even more closely (+0.86), come out the same way, at +0.36 or +0.04. Neither goes with a deeper drop.
The sharper check splits the owners 35 to 54 by whether they carry a mortgage, each counted as a share of all the area’s owner households. An owner with no mortgage has no cheap loan to give up, so under the lock-in effect the areas where more of them live ought to have kept selling. They didn’t. At similar price levels, the share of mortgage-free owners 35 to 54 lines up with a deeper drop (−0.39, or −0.28 with price adjusted the other way), as the share with a mortgage does (−0.33 or −0.51). Put the other way round, the fraction of an area’s owners 35 to 54 who have no mortgage goes with a smaller drop only weakly, if at all (−0.07 or +0.11). Among owners 65 and older, those with a mortgage and those without look alike, too (+0.20 and +0.29, or +0.40 and +0.44).
The age pattern itself turns up inside each part of the city we checked on its own. Among the 31 cheaper areas, where the median 2018–19 house sold for $30,000 to $235,000, the correlation is −0.71; among the 31 pricier ones, −0.44. In the 18 majority-Black areas it is −0.68, in the 17 majority-Latino areas −0.50, and in the other 27 −0.59. At similar prices, each 10 more points of owners 35 to 54 went with a sales drop about 9 points deeper.
Our theory: the lock is the next loan, not the old one
Here is our explanation, and it’s a theory, not a finding. The lock-in effect, as realtor.com describes it, has two parts: the cheap loan an owner gives up and the costlier one taken on. We think the second part is what held sales back in these areas, and the first added little. The area figures can’t show what any one owner did; they fit that version better than one built on the old loan.
Why would age carry it? Among Chicago owners 35 to 54, 78 percent carry a mortgage; among owners 65 and older, 59 percent own outright. Beyond that, what follows is our surmise about a possible mechanism, and none of these records measures it: an owner in the middle years who sells may often be buying again, perhaps something bigger, and would then need a loan at today’s rate on today’s prices, mortgage on the current house or not. An owner past 65 may more often make a move that needs no new loan, such as something smaller paid for with the proceeds, or a move in with family or into care. If so, higher rates would have raised the cost of the first kind of move far more than the second.
Borrowers tell a similar story
The federal mortgage records list each borrower’s age, and the Census Bureau counts Chicago households by the age of the householder, so purchase loans can be set against households of about the same age. It’s a rough comparison: the loan records use the main applicant’s age and the Census the householder’s, and those aren’t always the same person. In 2023–24 loans per household ran 29 percent below 2018–19 for ages 25 to 34 and 30 percent for 35 to 44, then 18 percent for 45 to 54, 15 percent for 55 to 64 and 13 percent for 65 to 74; households 75 and older took out 4 percent more. Under 25, a small group with 4 to 5 percent of the loans, the drop was 11 percent. In 2025 the spread widened: 39 percent fewer at 25 to 34 and 31 percent fewer at 35 to 44, about the same as before at 65 to 74, and 24 percent more at 75 and older. These are buyers who borrowed; a purchase paid in cash leaves no record here.
Other explanations we checked
Did older owners keep selling? Less than the table’s Change column suggests. In the 15 areas with the most owners 65 and older, the number of house sales by sellers whose names read as companies rose 8 percent from 2018–19 to 2023–24, from 1,936 to 2,092, and their share of those areas’ house sales went from 29 percent to 36 percent. Set company sellers aside and those areas lost 23 percent, not far from the 29 percent lost in the areas with the most owners 35 to 54. Across the city, companies sold 17.5 percent of houses in 2018 and 21.3 percent in 2024. The 35 to 54 link survives without them (−0.41, and −0.44 at similar prices); the link with owners 65 and older shrinks to +0.24. So the claim this page makes is about where sales fell furthest, not that older owners kept selling.
Is it estate sales? Heirs and executors would keep selling whatever the rate. In 2021 the deed records named an estate, executor or heirs as the seller on 3.2 percent of house sales in the 15 areas with the most owners 65 and older, against 1.2 percent in the 15 with the fewest, and those sales rose from 111 in 2021 to about 144 a year in 2023–24. That’s about 1 point of the difference between the groups. (The file began recording sellers’ names more fully in the autumn of 2020, so seller-type comparisons start in 2021.)
Is it company buyers? Their share of 2018–19 purchases adds nothing once price is counted (−0.06, or +0.20 with price adjusted the other way). Is it the pandemic buying spree, with recent buyers in no hurry to sell again? Areas where more owners moved in during 2020–2022 didn’t lose more sales at similar prices (+0.30 or +0.05). And is age standing in for the city’s racial geography? Holding fixed each area’s Black and Latino shares together with price, the share of owners 65 and older, 2020–21 refinances, recent movers and company buyers leaves the age link at −0.42.
What argues against it
These are figures for places, not people. The county’s file doesn’t record a seller’s age, so we can’t show that owners 35 to 54 in particular stopped selling, only that the community areas full of them lost the most sales. A household-level record of who sold and why could overturn the reading.
Something besides borrowing costs was already at work. At similar prices, the areas with more owners 35 to 54 were selling less in 2020, at record-low rates (−0.37). Measured against 2020–21 instead of 2018–19, the age link at similar prices weakens to −0.26, which leaves room for moves pulled forward into the pandemic years.
The mortgage split depends on the arithmetic. Adjusting for price one way, owners 35 to 54 without a mortgage line up with the drop slightly more closely than those with one; adjusting the other way, the mortgaged share lines up more closely. Both point the same direction, but how much the old loan matters isn’t settled by these numbers. And the borrower figures don’t single out the move-up years: the drop per household is as steep at 25 to 34 as at 35 to 44, so they can’t separate a renter priced out of a first purchase from an owner who decided not to trade up.
Some of the pattern is unexplained. The larger an area’s Latino share, the more sales it lost beyond what its owners’ ages and price level predict (−0.28 at similar prices and ages), and nothing in these files says why. The price adjustment is rough, one median per area, and the comparison stops in 2024 because the county keeps adding sales to its file for about 18 months after they close.
How we’ll test it
The check that can tell our theory from the textbook lock-in is the one about owners with no mortgage. They have no cheap loan to protect. If the old loan is what binds, areas full of them recover as soon as buyers do; if the next loan is what binds, they stay slow with everyone else who would have to borrow to move.
Sales, 2028. Once the county’s count of 2026 house sales is complete, we’ll take the first reading of the Assessor’s file on or after July 1, 2028 and measure 2026 house sales in the same 62 community areas against their 2018–19 average, with the same screens and company sellers included. The test variable is the one above: mortgage-free owner households whose householder is 35 to 54, as a share of all the area’s owner households, from the same 2020–2024 survey, compared at similar price levels both ways. If it lines up with lower sales at −0.15 or stronger under both adjustments, that’s consistent with our theory. If it reads −0.10 or higher under both, a positive reading included, while the mortgaged share reads −0.20 or stronger under both, that would count heavily against our theory and toward the old-loan version. Anything else leaves it open. In 2023–24 the two shares read −0.39 and −0.28 (mortgage-free) and −0.33 and −0.51 (mortgaged). On the 2025 sales recorded so far the mortgage-free share reads −0.31 and −0.21, and the mortgaged share −0.35 and −0.53; the completed 2025 count, due around the middle of 2027, will be a first look.
One looser expectation, which isn’t a test and can’t separate us from simpler explanations: if the 30-year rate falls below 6 percent and stays there for a year, the gap between the areas with the most and fewest owners 35 to 54 (19 points in 2023–24 and 21 so far in 2025; 11 points in 2023–24 with company sellers set aside) ought to start closing.
- Sales. Cook County Assessor parcel sales (dataset wvhk-k5uv), read September 30, 2026: single-parcel sales in Chicago’s eight assessment townships that pass the Assessor’s screens for arm’s-length sales (no sale of the same parcel within a year, nothing under $10,000, no deed types the Assessor excludes), house classes 202–210, 234, 278 and 295, placed in community areas by parcel. Buyers and sellers are sorted by the name on the deed, not verified: a company is a name containing LLC, Inc., Corp., Corporation, Ltd., LP, LLP, Company, Co., Properties, Holdings, Investments, Ventures, Capital, Development, Developers, Management, Realty, Group, Partners, Enterprises, Acquisitions, Equities or Estates, after banks and government bodies are set aside; an estate seller is a name containing Estate of, Executor, Administrator, Heirs or a similar probate term that isn’t a company. The 2025 figures are the sales recorded by September 30, 2026.
- Owners and households. Census Bureau American Community Survey 2020–2024 five-year estimates by census tract: mortgage status by age of householder (B25027), year the owner moved in (B25038), Hispanic origin and race (B03002). Majority-Black areas are those where non-Hispanic Black residents are at least half the population; majority-Latino areas, where Hispanic or Latino residents of any race are at least half. The two can’t overlap, because the Black count leaves out Hispanic residents; every other study area is in the third group. A tract belongs to the community area holding its Census interior point. Every owner share is of all the area’s owner households: owners 35 to 54 with a mortgage, and those without one, are each counted against that total. Households by age of householder (B25007, owners and renters) and the city’s 2025 mortgage share are one-year estimates for the City of Chicago.
- Loans. Home Mortgage Disclosure Act loan records from the FFIEC data browser, Cook County, originated loans: home purchase for 2018, 2019, 2023, 2024 and 2025, refinance and cash-out refinance for 2020 and 2021, owner-occupied, first lien, one to four units, excluding reverse mortgages, home-equity lines and business loans. Records through 2021 use 2010 census tracts, later ones 2020 tracts; both were placed in community areas the same way.
- Comparisons. Correlations are across the 62 community areas with at least 100 house sales in 2018–19. “At similar price levels” means a partial correlation that holds the area’s median 2018–19 house price fixed; we adjust for price two ways, on its logarithm and on the price itself, and give both where they differ. “Once age and price are both counted” adds the share of owners 35 to 54 as a second control. Each adjusted figure is a partial correlation: the measure and the change in sales are each regressed on the controls across the 62 areas, and the leftovers are correlated. The 10-point figure is the coefficient on the share of owners 35 to 54 in a regression of the change in sales on that share and the logarithm of price. Rates are yearly averages of Freddie Mac’s Primary Mortgage Market Survey, via the St. Louis Fed.
Computed by KCM Desk from the Cook County Assessor’s sales file (read September 30, 2026), federal Home Mortgage Disclosure Act records, the Census Bureau’s American Community Survey and Freddie Mac’s weekly rate survey (read November 18, 2026). Human-guided and edited — about this desk. If you spot an error, corrections come first.
Related Keep Chicagoland Moving coverage
- Chicago Housing, by the Numbers
- Chicago Houses That Closed in June Sold for 12 Percent More Than January’s in the Same Area, 2018 to 2024. Condominium Closings Ran Twice as High in May
- Who Is Buying Chicago? The Buyer’s Name on 245,849 Deeds: Companies Bought 13 Percent of Homes and 46 Percent of Those Resold Within Two Years
- What a Chicago Home Actually Sold For: 245,849 Deeds Since 2018, All 77 Community Areas
- Chicago’s January–April House Sales Now Read 5.6 Percent Below Last Year, Not 23 Percent Above — the County Added 1,147 Sales to 2025 After Our July Count

