Dallas County's housing market is splitting again, as it did in 2005Price tiers that pulled apart in the last cycle are pulling apart again, at about half the old speed

To compare the two cycles, the chart sets each era's tier indexes to zero at the quarter the bottom third peaked: the first quarter of 2005 in the last cycle and the second quarter of 2022 in this one. Dashed lines are 2005 onward; solid lines are 2022 onward. Scroll to step through the quarters.

The pattern repeats: in both eras the low tier turned first and the high tier kept going. What differs is the speed. 17 quarters after its peak, averaging the last two quarters, the current low tier is down about 9% and the gap between the top and bottom thirds, measured from the peak, is about 20%. The last cycle opened a gap that wide in 8 quarters, by Q1 2007. At the same 17-quarter mark, in Q2 2009, the earlier low tier was down 37% and the gap was 63%.

The high tier also behaved differently. In the last cycle it kept rising for nine quarters after the low tier peaked, then gave back most of that gain in 2008 and 2009 and ended 2012 near its 2005 level. In this cycle it has kept rising for 17 quarters and stands about 8% above its mid-2022 level, on the same two-quarter average.

Consumer prices rose 28% from the fourth quarter of 2020 to the third quarter of 2026. The dotted line on the nominal chart shows where a house would need to be to keep pace. The low tier's lead over that line narrowed every year after 2022 and was gone by Q1 2026. Switch to the adjusted view to see each tier in late-2020 dollars.

In late-2020 dollars, the high tier has gained about 16%, the mid tier about 1%, and the low tier has lost about 4%. An owner who bought a low-tier home in late 2020 has more dollars of value today and less purchasing power in that value than at purchase. An owner who bought in mid-2022 has neither.

The FHFA zip index does not yet show the current split. From 2021 to 2025 it rose 31% in the lowest-priced third of Dallas County zips and 34% in the highest.3 That index tracks only conventional loans bought by Fannie Mae and Freddie Mac, and only through 2025. The bottom-tier sales in our data increasingly use FHA financing or seller concessions, which the FHFA series does not capture.

We do not have loan data, so this is an estimate built from the sales records. For each FHA, VA and USDA purchase from January 2021 through June 2026 that has not been resold on the MLS, we assumed the program's minimum down payment with the upfront fee financed, a 30-year fixed rate at that month's Freddie Mac average, and scheduled payments since. We then marked the purchase price to the September 2026 level of its price tier's index. Individual homes scatter widely around an index, so rather than call each loan underwater or not, we used the spread the repeat-sales model measures between pairs of sales of the same house to estimate the probability that each home would sell today for less than the loan balance. Added up across 18,697 purchases, the expected number is 7,035, or 38%. Buyers who put down more than the program minimum owe less than the model assumes, so the figure is closer to a ceiling than a central estimate.

An FHA loan starts at about 98% of the purchase price once the upfront premium is financed, and a VA loan with the funding fee financed starts above 100%, so a recent buyer in either program has little room for a price decline or for an unlucky sale. The tier's direction decides how much worse it gets. In the low tier 43% of purchases, about 3,900 homes, are expected to sell for less than the balance, against 34% in the mid tier and 28% in the high tier. Low-tier buyers from 2023 and 2024 are the most exposed, at 57% and 56%. Buyers from 2021 bought before the run-up ended and nearly all have equity.

Most underwater borrowers are close to even. Of the 7,049 low-tier purchases from 2022 onward, about 3,400 are expected to be above 100% loan-to-value and about 1,700 above 110%. That is a smaller hole than in 2009. It still matters, because a borrower without equity cannot sell to pay off the loan. Selling costs in Dallas County, with commissions and the concessions now common at the bottom of the market, run to roughly 8% of price. Counting those costs, about 69% of 2023 buyers and 71% of 2024 buyers would need to bring cash to closing, so a job loss, divorce or move can become a short sale, a loss-mitigation case or a foreclosure.

The 2005 split became a crash in the low tier because foreclosures put a steady stream of homes on the market at whatever price cleared them. Today's owners mostly have fixed-rate loans with verified income behind them, and the bank-owned share of sales is a small fraction of what it was. That makes a repeat of 2008 to 2011 in the bottom tier unlikely without a large rise in unemployment.

Sellers who need to move cut their price or pay the buyer's costs. Owners who do not need to move stay put. Fewer homes trade, and those that do clear a little lower each year, in nominal or in real terms. The high tier in the last cycle shows what that looks like over a long stretch: it ended 2012 about 5 points above its early-2005 level in nominal terms and 11% below its 2004 value after inflation, with no crash along the way.

Where losses would land

The losses from a slow slide fall on different parties than in 2008, when banks and private mortgage securities absorbed much of the damage. FHA loans are insured by the government through FHA's Mutual Mortgage Insurance Fund, and VA loans are partly guaranteed; the Mortgage Bankers Association put the FHA delinquency rate at 11.79% in the second quarter of 2026, against 2.72% for conventional loans.11 ICE reported early this year that FHA loans accounted for about 80% of the rise in seriously delinquent and foreclosed mortgages, and FHA's pandemic-era loss mitigation options expired on Sept. 30, 2025.1213 Builders carry the cost of the buydowns and concessions they use to keep selling. Texas counties, cities and school districts depend on property taxes set from appraised values, so a long decline at the bottom of the market shows up in their budgets. Owners in the bottom tier, whose home is often most of what they own, lose the equity they would have used to trade up, retire or weather an emergency.

The bottom-tier zips with the most FHA, VA and USDA purchases from 2023 onward, in low and mid tiers combined, are 75149 (628), 75115 (597), 75181 (586), 75241 (557), 75043 (547), 75216 (532). They are in Mesquite, DeSoto, Garland and southern Dallas. The equity estimate is not marked zip by zip, so these counts show where the loans are, not which zips have fallen most. They are loan-volume counts, not a risk ranking of neighborhoods, and they are no basis for deciding where to lend or how to price a loan.