We matched 2,329,694 Florida homes to their own earlier sale and asked one question of each: after the roughly 8% it costs to buy and sell, did the owner get back more than they paid? Grouped by the year they bought, the answer runs from 4.9% to 85.8%.
The upper line counts owners who netted less than they paid once buying and selling costs are subtracted. The lower line counts those who sold below their purchase price outright. The space between them is what transaction costs alone did to people whose houses did not fall in value.
2006 was the worst year in living memory to buy a Florida house. Of the 129,350 homes bought that year that have since sold, 85.8% came back for less than the owner paid once costs are counted. 81.4% lost money outright, before a single fee.
2020 was the safest year in the series at 4.4%, with 2012 just behind at 4.9% and 2011 at 5.4%. Two very different moments produced almost the same result: buying at the bottom of the crash, and buying in the last quiet months before the run-up.
Then it turned fast. 2021 buyers are at 12.8%. 2022 jumps to 48.9% and 2023 to 55.5%. On this measure, buying in 2023 has so far been worse than buying in 2004, and only 2005, 2006 and 2007 were worse.
The gap between the two lines is its own finding. In 2023, 28.7% sold below what they paid but 55.5% lost money once costs came out. For roughly one buyer in four, the house held its value and the transaction still cost them.
101,213 matched pairs from buyers in those two years who have already sold. Statewide 51.3% lost money after costs and 26.2% sold below what they paid. St. Johns and Flagler are the two hardest-hit counties in the state, both near 71%. They are close enough (70.6% and 70.3%) that we do not rank one above the other: the gap between them moves depending on how properties listed in more than one feed are matched, so treat them as tied. Among the largest markets, Miami-Dade is the most resilient at 31.7%. Counties with fewer than 300 matched pairs are not shown.
| County | Matched pairs | Lost after 8% costs | Sold below purchase |
|---|---|---|---|
| St. Johns | 1,954 | 70.6% | 39.2% |
| Flagler | 1,023 | 70.3% | 39.0% |
| Charlotte | 1,672 | 66.4% | 50.2% |
| Osceola | 2,853 | 66.1% | 39.9% |
| Sarasota | 3,556 | 64.9% | 46.0% |
| Manatee | 3,120 | 64.0% | 42.7% |
| Sumter | 586 | 63.3% | 37.5% |
| Walton | 1,154 | 61.2% | 34.5% |
| Indian River | 382 | 60.2% | 34.6% |
| St. Lucie | 2,022 | 59.4% | 31.3% |
| Polk | 4,127 | 58.8% | 31.5% |
| Lake | 2,340 | 58.4% | 25.3% |
| Okaloosa | 1,081 | 58.0% | 25.5% |
| Volusia | 2,619 | 56.1% | 30.0% |
| Brevard | 3,503 | 55.9% | 26.8% |
| Clay | 1,059 | 55.3% | 25.6% |
| Santa Rosa | 1,073 | 55.2% | 18.8% |
| Pasco | 2,487 | 55.2% | 28.1% |
| Marion | 2,859 | 54.7% | 27.7% |
| Nassau | 500 | 54.2% | 19.6% |
| Martin | 580 | 54.1% | 28.6% |
| Escambia | 1,920 | 54.0% | 24.4% |
| Hillsborough | 6,890 | 53.8% | 27.9% |
| Duval | 4,393 | 52.9% | 26.2% |
| Hernando | 1,152 | 52.8% | 26.2% |
| Monroe | 685 | 52.3% | 26.1% |
| Alachua | 1,168 | 52.1% | 18.9% |
| Citrus | 317 | 50.8% | 25.9% |
| Seminole | 1,929 | 49.8% | 18.5% |
| Pinellas | 7,231 | 49.0% | 27.0% |
| Orange | 5,221 | 47.8% | 18.0% |
| Palm Beach | 9,979 | 43.3% | 21.3% |
| Broward | 9,847 | 42.0% | 19.0% |
| Miami-Dade | 8,547 | 31.7% | 12.5% |
| Putnam | 327 | 26.0% | 10.4% |
The clearest pattern in the recent cohort is not where a home is but when it was built. Among 2022 and 2023 buyers who have already sold, newly built stock is roughly thirty points worse than pre-2000 stock.
| When the home was built | Matched pairs | Lost after 8% costs | Share of cohort |
|---|---|---|---|
| Built 2020 or later | 13,193 | 65.7% | 13.0% |
| Built 2010 to 2019 | 8,949 | 73.5% | 8.8% |
| Built 2000 to 2009 | 17,163 | 60.3% | 17.0% |
| Built before 2000 | 61,887 | 42.6% | 61.2% |
The worst band is not the newest. Homes built between 2010 and 2019 lost money more often than homes built since 2020, which is what you would expect if the problem is competing with a builder: they sit in the same subdivisions as active new inventory, are priced against it, and carry none of the premium a brand new house still commands.
Splitting it by the year of purchase shows the pattern is not a quirk of one cohort. From 2021 onward the 2010s column is the worst in every row.
| Bought in | Built 2020+ | Built 2010s | Built 2000s | Built pre-2000 |
|---|---|---|---|---|
| 2019 | 11.9% n=774 | 13.8% n=17,836 | 8.7% n=22,144 | 6.5% n=65,617 |
| 2020 | 6.8% n=7,994 | 5.6% n=10,787 | 3.9% n=18,896 | 4.0% n=55,021 |
| 2021 | 16.8% n=8,403 | 18.0% n=10,881 | 13.6% n=19,788 | 10.9% n=57,927 |
| 2022 | 59.1% n=6,963 | 70.8% n=6,009 | 57.6% n=11,502 | 41.3% n=39,494 |
| 2023 | 73.2% n=6,230 | 79.1% n=2,940 | 66.0% n=5,661 | 44.8% n=22,393 |
A 2022 buyer of a 2010s home is at 70.8%. A 2022 buyer of a pre-2000 home in the same market and the same year is at 41.3%. That is a 30 point spread produced by nothing but the age of the house. The 2023 row runs higher still, but that cohort is younger and skews toward people who had to sell quickly, so the 2022 row is the fairer comparison.
It is not a permanent rule. In 2020 the newest stock was the worst performer and the ordering inverts. What the table describes is this run-up and this correction.
Two different things can make a county look bad: having more new construction in its cohort, or being worse at every vintage. Standardising separates them. The fourth column is what each county's rate would be if its cohort had the statewide mix of build eras, so the drop from the actual figure is the part attributable to composition alone. Pair counts here are slightly lower than in the county table above because a small share of records carry no build year and are left out of this comparison rather than guessed at.
| County | Pairs | Actual | With the state build mix | Explained by mix | Built 2020+ |
|---|---|---|---|---|---|
| St. Johns | 1,954 | 70.6% | 60.8% | 9.8 pts | 38.3% |
| Flagler | 1,023 | 70.3% | 66.9% | 3.4 pts | 24.2% |
| Charlotte | 1,672 | 66.4% | 66.9% | -0.5 pts | 16.7% |
| Osceola | 2,853 | 66.1% | 56.8% | 9.3 pts | 22.3% |
| Sarasota | 3,544 | 65.0% | 65.2% | -0.3 pts | 15.9% |
| Manatee | 3,118 | 64.0% | 59.6% | 4.4 pts | 24.7% |
| Walton | 1,154 | 61.2% | 55.8% | 5.4 pts | 33.8% |
| St. Lucie | 2,022 | 59.4% | 57.5% | 2.0 pts | 19.3% |
| Polk | 4,127 | 58.8% | 54.4% | 4.4 pts | 19.1% |
| Lake | 2,340 | 58.4% | 54.3% | 4.1 pts | 20.9% |
| Okaloosa | 1,081 | 58.0% | 56.9% | 1.1 pts | 18.2% |
| Volusia | 2,619 | 56.1% | 58.5% | -2.4 pts | 9.2% |
| Brevard | 3,503 | 55.9% | 57.0% | -1.1 pts | 13.4% |
| Clay | 1,059 | 55.3% | 50.7% | 4.7 pts | 19.5% |
St. Johns is the clearest case of the first kind: 9.8 points of its gap above the statewide 51.3% come from composition, because 38.3% of its cohort was built since 2020 against 13.0% statewide. Flagler, Charlotte and Sarasota are the second kind. Composition explains almost none of their gap and they are underwater at every vintage, which is a market that ran up and came back down rather than a market competing with a builder.
Combining 2022 and 2023 buyers to keep the samples usable, the vintage gap holds inside individual counties. In the hardest-hit dozen, between eight and nine in ten buyers of 2010s-built homes lost money, against roughly four to six in ten for pre-2000 homes in the same county and the same two years.
| County | Built 2010s | Built 2000s | Built pre-2000 | Spread |
|---|---|---|---|---|
| Sarasota | 87.0% n=315 | 79.4% n=544 | 59.1% n=2,123 | 28 pts |
| Polk | 85.5% n=427 | 68.7% n=926 | 41.0% n=1,984 | 44 pts |
| Charlotte | 85.2% n=115 | 82.0% n=261 | 60.1% n=1,017 | 25 pts |
| St. Johns | 85.0% n=367 | 68.3% n=369 | 51.9% n=470 | 33 pts |
| St. Lucie | 84.9% n=152 | 69.0% n=467 | 48.2% n=1,013 | 37 pts |
| Manatee | 84.0% n=494 | 70.8% n=510 | 50.8% n=1,344 | 33 pts |
| Walton | 82.2% n=213 | 65.0% n=274 | 49.5% n=277 | 33 pts |
| Hillsborough | 82.0% n=762 | 68.8% n=1,213 | 40.8% n=3,776 | 41 pts |
| Osceola | 81.7% n=639 | 65.7% n=877 | 47.6% n=701 | 34 pts |
| Pasco | 80.5% n=200 | 68.1% n=383 | 41.4% n=1,359 | 39 pts |
| Clay | 79.1% n=115 | 62.2% n=254 | 38.6% n=484 | 40 pts |
| Marion | 78.0% n=200 | 60.2% n=573 | 42.2% n=1,447 | 36 pts |
The counties at the top are closely bunched and we do not rank them: the leaders sit within a few points of each other on samples of a few hundred, which is not enough to separate them. The spread column is the more reliable signal, and it is widest in Polk and Hillsborough rather than in the counties with the highest headline rates.
This is association, not cause. Build year stands in for buying new from a builder; we do not observe who the seller was on the earlier sale. Newer subdivisions also differ in ways this does not separate, notably community development district assessments and insurance costs, each of which reaches resale value through its own channel. What the numbers support is that newer housing has performed worse, and that in some counties the amount of it explains a large part of the gap. They do not establish that builder competition is the mechanism.
Years marked partial are incomplete cohorts: only the faster sellers have returned to market so far, and faster sellers lose money more often, so those rows will move.
| Bought in | Matched pairs | Lost after 8% costs | Sold below purchase |
|---|---|---|---|
| 2026 · partial | 1,221 | 31.9% | 15.6% |
| 2025 · partial | 9,265 | 38.8% | 18.7% |
| 2024 · partial | 19,450 | 52.8% | 29.1% |
| 2023 | 37,227 | 55.5% | 28.7% |
| 2022 | 63,986 | 48.9% | 24.7% |
| 2021 | 97,026 | 12.8% | 4.3% |
| 2020 | 92,716 | 4.4% | 1.2% |
| 2019 | 106,381 | 8.2% | 2.0% |
| 2018 | 116,995 | 12.0% | 2.8% |
| 2017 | 126,230 | 11.6% | 2.8% |
| 2016 | 135,553 | 10.4% | 2.8% |
| 2015 | 145,864 | 9.5% | 3.0% |
| 2014 | 140,085 | 9.1% | 3.0% |
| 2013 | 138,177 | 7.9% | 2.6% |
| 2012 | 129,981 | 4.9% | 1.6% |
| 2011 | 122,973 | 5.4% | 2.2% |
| 2010 | 113,323 | 10.3% | 5.1% |
| 2009 | 101,338 | 17.1% | 10.1% |
| 2008 | 75,653 | 43.6% | 33.2% |
| 2007 | 87,111 | 76.0% | 68.8% |
| 2006 | 129,350 | 85.8% | 81.4% |
| 2005 | 137,137 | 76.0% | 69.8% |
| 2004 | 87,475 | 51.5% | 43.9% |
This is not a claim about everyone who bought in a given year. It counts only homes that have sold again, so owners who bought in 2006 and never left do not appear, and on the whole they did better than the ones who sold. It is also not a forecast: the 2022 and 2023 figures describe the portion of those cohorts that has already resold, which skews toward people who had to move quickly. Southwest Florida is thin in the recent cohort because of feed coverage, not because nothing happened there, and counties below the reporting floor are left out rather than estimated.
Source: Momentum's licensed MLS closed-sale archives across eight Florida feeds, roughly seven million closings running back to 2001. A property is matched to its own earlier sale by county and parcel number, which is present on 99.6% of records. For each consecutive pair of closings the earlier is the purchase and the later the resale. Net proceeds are the resale price times 0.92, an 8% round trip covering commission, doc stamps, title and closing costs at both ends. Underwater means those proceeds fall below the purchase price. Nominal dollars, no inflation adjustment. Excluded: non-residential property types, closings under $10,000, and records with no usable parcel number. The series starts in 2004 because feed composition before then is uneven and not comparable across years; from 2004 the two largest feeds hold a steady 28–38% and 47–53% share of every year. Aggregates only; no individual transaction is disclosed. Deemed reliable but not guaranteed. Equal Housing Opportunity.
Listing failures by county, 2003 to 2025 · The monthly listing reality report · Price-cut tracker
