What counts as a stalled construction site

Every submarket has one. A site that was framed to the third floor in March and still looks exactly the same in September. Before you flag it in a market report or mention it on an earnings call prep doc, it helps to have an actual definition rather than a gut feeling from driving past it twice.

A working definition

Most analysts land on something close to this: a site counts as stalled when there has been no visible change in construction stage for two consecutive quarterly observations, and that gap isn't explained by a known, dated cause like a weather season or a scheduled inspection hold. One quiet quarter is normal. Steel sits, crews rotate to another job, a permit amendment takes six weeks. Two quiet quarters in a row, with no crane movement, no new material staging, no change in the slab or framing line, is the point where "slow" becomes "stalled."

The stage matters as much as the duration. A site sitting at excavation for six months reads differently than a site sitting at topped-out-but-unclad for six months. The first could be a land banking play or a slow entitlement process that hasn't broken ground in earnest yet. The second usually means the money ran out mid-build, which is a much more interesting data point for a REIT tracking competitive supply. A vertical shell with no cladding, no MEP rough-in progress, and tarped openings going into a second quarter is about as clean a stall signal as this asset class produces.

Context rules a lot of this out, too. Seasonal shutdowns in cold-weather markets aren't stalls. A site waiting on a utility connection that's a known municipal bottleneck isn't a stall, it's a queue. What matters is whether the project's own capital or approvals have hit a wall.

Pause versus dead deal

This is where most internal trackers get sloppy, because both look identical from the street for a while. A paused project still has a path back. The entitlements are current, the GC hasn't demobilized equipment off-site, and there's often a reason you can find in a permit filing or a lender's public disclosure, a refinancing, a change order dispute, a tenant pre-lease that fell through and the sponsor is re-tooling the program. A dead deal shows different signs over time: equipment leaves the site rather than sitting idle, fencing comes down or gets repurposed, the site starts getting used for parking or staging for an unrelated job, or vegetation starts coming up through gravel that used to be a laydown yard.

A single observation rarely tells the difference. You need the trend. A site unchanged for one quarter is just unchanged. A site unchanged for three or four quarters running, especially with equipment count dropping between each observation, is a much stronger candidate for dead than paused. That's the case for tracking on a fixed cadence instead of checking in whenever a site happens to catch your eye. A one-off site visit tells you the state; a quarterly series tells you the trajectory, and trajectory belongs in a supply-pipeline model.

For a submarket with a dozen or more live permits, keeping that quarter-over-quarter comparison current by eye, or by driving the sites yourself, gets unworkable fast once you're covering more than one or two assets. Construction Pipeline builds that comparison as a quarterly status table per site, so a stall shows up as the same stage reading two quarters running instead of a memory of what a site looked like last time you drove past it.

If competing supply in your submarket is something you're tracking quarter to quarter already, it's worth seeing what a standing status table looks like for your sites.