The 127% August gain was a merger; asset sales fell 21%

August's U.S. commercial real estate sales total of $107 billion was up 127% from a year earlier, and most of that gain belongs to the stock market: MSCI attributes $70 billion of the month to merger and acquisition activity, leaving about $37 billion of asset-level trades, down roughly 21% against August 2025. Anyone financing a building should be reading the second number.

The merger doing the lifting was AvalonBay Communities' combination with Equity Residential, the apartment company with an enterprise value around $70 billion now trading as Vivmark Residential, which MSCI names as the main driver of August volume and the reason multifamily transaction volume rose 402% year over year. That is a corporate combination recording as property volume, and it prices what a public buyer paid for a portfolio it will run in-house, not what a private buyer paid for a single asset it had to finance at August's cost of debt. We made the point in August when the combined company began trading, and August's final data has not changed it.

Outside that deal, the sector detail is narrow: industrial was the only traditional property type to grow, up 14% to $11.5 billion, with senior housing close behind at 8%, while every other class slipped—no data-center assets traded in August at all, hotel volume fell 45%, and office, retail and development sites each landed below their year-earlier levels. MSCI's analysts read the month as deep liquidity paired with a decelerating pipeline.

Prices barely moved. The RCA CPPI U.S. National All-Property Index was up 0.1% from a year earlier, and the average cap rate across all August transactions was 6.01%, down 80 basis points. A flat price index and a falling blended yield in the same month point to composition rather than repricing: when $70 billion of a $107 billion month is one low-cap-rate apartment merger, the blend drops without saying anything about what a 1990s office building or a grocery-anchored strip center is worth. Sponsors marking portfolios to that 6.01% are marking to a merger.

JPMorgan's capital markets team, in a note to investors, argues that the jump in yields at both the long and short ends arrived late in the third quarter, so deals already in motion were likely unaffected. The bank expects third-quarter volume to land roughly in line with expectations, and the reading is that higher debt service has slowed decision-making rather than stopped it. The specific caution is multifamily, where cap rates already sit low enough that costlier debt could kill trades.

That caution lands on the same sector the merger flattered, and the apartment bid has split into an income half and a scarcity half, with value-add buyers setting the clearing basis lower while patient capital underwrites the 2028-29 supply gap; August's 402% belongs to neither half. The month establishes that asset-level trades across all property types are down 21%, and if apartment single-asset trades are clearing on weaker economics than the headline implies, the financing math for value-add buyers gets harder from here, not easier.

Industrial's 14% gain carries its own weight: industrial pricing is now a rents-and-scarcity trade in which the operator, the land basis and the data-center pull set the clearing price, and the building does less of the work than it used to. A month in which industrial was the only traditional sector to add volume, against a data-center class that set records in July and then traded nothing, is what that position predicts.

July's $74.4 billion total, lifted by BlackRock's $33.7 billion acquisition of Aligned Data Centers, was the two-decade high our own coverage flagged, and it was a month propped up by a single transaction. August's data-center volume was zero. That swing says less about demand for server capacity than about how episodic price discovery in the class remains: there is no continuous bid for data centers, only offerings, and each one sets its own comps. The live test is Mapletree's 22-building, 3.1 million-square-foot portfolio, which will show whether buyers pay for stabilized income or keep pricing the class off build-out cost.

For the year, volume through August is $483 billion, up 53%, but portfolio and entity-level deals are up 221% against a 12% gain in single-asset sales. The same composition problem repeats at scale, and growth is arriving through consolidation rather than through buildings changing hands—that matters most to the people who need comps. A lender sizing a loan or a sponsor marking a portfolio pulls from the single-asset series, which is expanding at less than a quarter of the headline rate.

For boards weighing an exit, the practical read is that a sponsor with a portfolio large enough to be a platform has a buyer base that a single-asset seller does not, which suggests the market is paying for scale and control while paying ordinary prices for buildings. Entity-level processes are where the volume and the premium sit, and the 221% figure is the evidence.

September is the first month whose deals were decided with the higher long end in the room, so it is the cleanest available test of whether capital costs are delaying transactions or merely rescheduling them. If the non-merger line recovers toward flat while the headline holds, the market is broader than August suggested. If the headline holds and the non-merger line does not, the volume belongs to corporate finance, and the $37 billion of asset-level trades, down 21%, is the figure that matters to anyone whose collateral is a building.