Multifamily's coming shortage is a financing problem, not a demand one

The apartment construction pipeline has fallen to roughly 475,000 units, about 3.5 percent of existing inventory and the lowest level Cushman & Wakefield has recorded since 2013, which Sean Rae, senior managing director of capital markets at Crow Holdings Development, reads as the setup for the next cycle rather than a warning about this one. Supply should keep tightening into 2028 and 2029, he told IREI, and muted starts make a shortage several years out the likely outcome, though the squeeze will not arrive uniformly across markets, and that variance is where the underwriting work now sits.

Demand is not doing the work. RealPage reported apartment occupancy improved to 95.5 percent in the second quarter of 2026, with rents still slightly below year-earlier levels and concessions elevated, so occupancy climbing while pricing power stays flat is what a market looks like when it heals from the supply side. Every apartment model being marked this fall carries that distinction: the rent growth in those models depends on a delivery calendar, not a demand forecast, and Rae's 2028-2029 window implies that calendar is longer than a standard five-year hold, which puts the pipeline number in the same conversation as the debt stack rather than the rent comps.

Developers have been making this bet for a year, ever since Hines pivoted from buying to building on the logic that a construction freeze across markets produced a scarcity advantage acquisitions could not match; the apartment version of that argument has now moved from thesis to data, because a 3.5 percent pipeline is a count rather than a projection. What has not been settled is who can afford to hold it.

The apartment bid has split in two: loan-carrying basis trades on standing product and below-replacement land bets on ground-up, with the middle failing to clear. Merchant building that needs a stabilized exit at a cap rate the market no longer pays does not get repaired by a thin pipeline; what a thin pipeline does is raise the odds on the land-bet bucket and raise the price of patience. Rae's account of financing new development in this environment makes the practical stakes plain, because the rescue capital that has kept maturing apartment debt out of distress, the extensions and preferred equity this publication has tracked, is built for owners of standing buildings rather than sponsors borrowing against a 2029 delivery. Construction capital, not land, is the scarce input.

Watch the starts line through 2027. If it stays muted, leases signed into the 2029 delivery years get negotiated against the thinnest pipeline since 2013, and concessions will move before asking rents do.