Answer in brief
CoStar’s 28 September release shows opposing price movements across two weightings of US commercial sales. Fewer August deals and projected supply reductions complicate the recovery picture.
Two readings of the same transaction market
US commercial property prices moved in opposite directions in CoStar’s September release, dated 28 September 2026 and covering sales through August. Its value-weighted composite index fell 1.3% during the month, while its equal-weighted index rose 1.4%. The split is the main news: a statement that commercial property simply rose or fell would discard information about which transactions drive each measure and how the market is being summarised.
The value-weighted series gives greater influence to expensive transactions common in core markets. The equal-weighted measure is more affected by numerous lower-priced transactions typical of secondary and tertiary markets. These are different statistical lenses, not two incompatible estimates of one average selling price. Their divergence suggests that an account of recovery should identify the asset population before drawing a conclusion about the whole market.
The annual comparison also diverges
The value-weighted index reached 236 and stood 1.4% below August 2025. The equal-weighted index reached 318 and was 2.2% above a year earlier. Their index levels are not dollar prices, and dividing 318 by 236 does not reveal a premium available on a building. For interpretation, the changes within each series matter more than the numerical gap between their separately constructed index levels.
The monthly decline was the value-weighted measure’s fifth in succession. The equal-weighted increase was its first since March. CoStar also put the former 19.4% below its July 2022 peak and the latter just 0.5% below its March 2026 high. The peaks occurred at different times, so the distances describe each series’ history rather than a directly matched comparison of identical property portfolios.
| Measure | August result | Annual comparison |
|---|---|---|
| Value-weighted composite | 236; −1.3% monthly | −1.4% |
| Equal-weighted composite | 318; +1.4% monthly | +2.2% |
| Repeat-sale count | 1,534 pairs | −8.6% |
| Repeat-sale consideration | $11.3bn | −10.3% |
| Distressed repeat sales | 45 pairs; 2.9% of sample | Not supplied here |
August trading cooled despite a stronger annual total
CoStar counted 1,534 repeat-sale pairs in August, generating $11.3 billion of consideration. The transaction count was 8.6% lower than a year earlier, while dollar volume fell 10.3%. Those are monthly observations within the repeat-sales sample. They do not measure every transaction in the country, nor the value of buildings that remained unsold, and they should not be presented as a census of the whole commercial stock.
Over the twelve months ending in August, however, pair volume reached $161 billion, up 17.2%. This is another timing distinction: a strong trailing total can coexist with a weaker latest month. The editorial implication is to check both transaction prices and trading activity. A rising index in one part of the sample does not, on its own, establish that liquidity is improving at the same time.
Distress is concentrated within the sample
The report identifies 45 distressed August sales, or 2.9% of all repeat-sale pairs. Within the investment-grade segment, 22 distressed transactions represented 9% of 244 pairs. General commercial properties accounted for 23 distressed trades among 1,290 pairs, or 1.8%. This difference is useful because the aggregate percentage conceals a much higher observed distressed share in the smaller, more expensive transaction cohort.
The denominator still matters. A share of completed repeat sales cannot reveal how many owners are struggling but have not sold. Nor does a distressed sale establish the condition of every neighbouring building. The data show the character of transactions reaching this particular sample. Applying the same percentage to the entire commercial inventory would replace a measured trading statistic with an unsupported estimate of financial vulnerability.
Construction forecasts are a separate evidence category
CoStar projects quarterly deliveries ending in September to be 63% below the fourth-quarter 2023 cycle peak. It also projects 482 million square feet of office, retail and industrial deliveries over the twelve months ending in September, down 19.5% from the comparable period. These are the producer’s forward estimates in this release. They should not be promoted into confirmed September outcomes simply because publication occurs near quarter-end.
Earlier financing context comes from the Federal Reserve’s July loan-officer survey, released on 3 August and generally covering the second quarter. Banks reported easier standards for some commercial-property lending, while construction and land-development demand weakened on balance. The survey concerns lenders’ reported conditions, not August sale prices. It helps frame financing constraints without independently validating CoStar’s transaction sample or proving what caused the observed price split.
Match the benchmark to the decision
Our interpretation is that market breadth and capital value are recovering unevenly. Before using a national index, a reader should identify whether the property resembles the high-value cohort or the more numerous general-commercial trades, then examine local leases, occupancy and actual comparable sales. The report does not supply the operating income or financing terms needed to value an individual building or promise an investment return.
The next useful evidence will be another comparable month of repeat sales and the replacement of projected supply figures with observed results. Until then, the defensible conclusion is a divergence between price measures alongside weaker August trading volume. That combination preserves the information in the release: different asset groups can move differently, and a forecast of less construction is not yet proof of stronger tenant demand.
Questions and answers
How can both price indices be correct?
They assign different influence to transactions. Expensive properties matter more in the value-weighted measure; numerous lower-priced sales have greater influence in the equal-weighted view.
Does a 2.9% distressed share describe every building?
No. It describes 45 distressed transactions among the 1,534 repeat-sale pairs in August. It cannot be treated as the share of the entire building stock under financial stress.
Has the projected delivery decline already been measured?
The report labels its September delivery and absorption figures as projections. Those estimates must remain separate from the August transactions used in the price indices.