Real Estate Outlook 2026

Builders in the May 1 print were sitting on 10.3 months of supply, the top of a 52-week range whose floor is 7.8, and up 10.75% in a month. That is the number defining this market, not the mortgage rate. The 30-year mortgage was 6.49% on July 9 and has gone almost nowhere: up 0.15% over 30 days, up 1.88% over 90, inside a 52-week band of 5.98 to 6.75. Nothing about that says shock. It says stalemate.

Volume is where the adjustment is happening. New home sales in the May 1 print ran at 580K against a 52-week high of 748 and a low of 576, down 7.35% over the prior 30 days. Housing starts printed 1177K, the 52-week low outright, down 15.45% over 30 days and 15.02% over 90. Builders are not cutting price into the inventory, they are cutting the inventory. Total completions at 1313K, down 8.12% over 30 days and also at the bottom of their 52-week range, say the same thing from the far end of the pipeline.

Sentiment fits, though the table gives a level and no history: the NAHB index read 34 on April 30, with buyer traffic at 22. Both are April vintages, so read them as a snapshot rather than a trend. An index below 50 means more builders call conditions poor than good, and a traffic reading of 22 describes showrooms nobody is walking into.

The freeze is not a rate event. It is a price-expectation event: owners who financed cheap will not mark down, buyers at 6.49% will not pay up, and the market settles the difference by doing less business.

Multifamily has already given up. Starts of 5+ unit buildings printed 284K in the May 1 vintage, down 41.56% over 30 days and 39.45% over 90, and that 284K is the low of a 52-week range whose high is 486. Permits in the same category held up far better at 468, off 4.68% in a month against a 561 high. Developers still hold the paper. They are not breaking ground on it.

Single-family is a different animal, and the contrast matters. Single-family starts at 882K fell only 1.89% over 30 days, and single-family permits rose 1.25% to 892. The pain is concentrated in the part of the market financed like a bond rather than the part financed like a house. An apartment project is underwritten against a cap rate, that cap rate has to clear a risk-free rate that keeps rising, and when it does not clear, the project simply does not start.

The vacancy backdrop is a January vintage and belongs in the background rather than the news: 15305 thousand vacant units, the top of a 52-week range whose low is 15039. There is no shortage of empty roofs in aggregate. There is a shortage of ones people can carry at 6.49%.

A 41.56% collapse in multifamily starts, with total completions already at their 52-week low of 1313K, is a rental supply cliff building for 2027 and beyond, because apartment buildings deliver years after the shovel goes in. The construction cut that looks like discipline this year is the mechanism that manufactures the next rent shock. Cheap for landlords who own today. Expensive for the inflation print a Fed stuck on hold will have to answer for.

XLRE at 44.45 has done very little: down 0.51% on the week, down 1.05% over 30 days, up 3.81% over 90, and below its 52-week high of 45.65. Homebuilders have done something. XHB fell 3.51% in a week to 108.61, handing back part of a 4.59% gain over 90 days, with its 30-day change still positive at 0.72%. Financials, the lender side of the identical trade, are up 9.73% over 90 days and 5.87% over 30, at 55.71.

XLF and XLRE were both priced on July 11, so the 90-day comparison is a fair one: banks have beaten property by a wide margin over the same window. Read literally, that is the market saying the credit is fine and the equity is not. Loan books get paid. Builders and developers eat the volume decline.

The tape has this half right, at best. The series that would settle it, CRE loan delinquency at banks, is not in our table, and I am not going to invent it. What the table does show is a lender complex the market has bid hard and a builder complex it has been selling for a week, inside a regime our own macro state gives an equal chance of resolving into stagflation as into a soft landing. Only one of those two paths gets the mortgage rate down, and it is not the one where real yields keep climbing.

The REIT signal is the subtle one. XLRE prices duration risk and credit risk in the same instrument, and it has held a 3.81% 90-day gain while going nowhere for a month. That is not a market braced for a commercial property accident. That is a market that has decided the rate path is the only question and has not yet been given the answer.

Case-Shiller printed 332.678 on April 1, its 52-week high against a low of 326.695, up 1.83% over 90 days. The print is 102 days old and it is the most backward-looking number on this page, an average of closings negotiated earlier still. Prices are the last thing holding in residential real estate, and they are holding on the oldest data available.

They give way before mortgage rates do. Every forward-looking series in the table has already broken: months supply at 10.3 is the 52-week high, starts at 1177K are the 52-week low, new home sales at 580K sit a whisker above their 576 low, and buyer traffic was 22 when it was last measured in April. A builder carrying 10.3 months of unsold inventory eventually discounts, that discount reaches a price index two quarters later, and the record Case-Shiller level is therefore describing a market that has since stopped existing.

The case against that is real, and it lives in the mortgage rate. At 6.49% the 30-year sits inside a 52-week range with a floor of 5.98, and it has been drifting the wrong way for buyers, up 1.88% over 90 days. A soft inflation print that pulls the rate toward that floor would meet a market whose starts are already at a 52-week low, and prices would clear rather than crack. Single-family permits ticking up 1.25% in the May 1 print says builders are keeping that option alive. XHB up 0.72% over 30 days says equity holders are paying for it.

Watch the next months-supply print above everything else. If 10.3 keeps climbing while Case-Shiller is still making highs, the two series are describing different years, and the newer one wins.