The LA Cost Ceiling Didn't Break. It Just Moved.

For the last three years, every trade publication and every lender's quarterly letter has been waiting for LA construction costs to come down. They have not come down. They have rotated. That distinction matters, because the investors and operators who keep underwriting a 2027 reset are, in effect, underwriting a reversion to a market that no longer exists.

I can only write what I can see from my own desk, and my desk sees five active jobs: a forty-three-unit ground-up on Fairfax, a twelve-unit repositioning in Mid-Wilshire, two gut renovations in Hancock Park, and a structural retrofit on an older courtyard building in Beverly Grove. What follows is the ledger those five jobs hand me, read honestly.

The ceiling that didn't break.

In early 2023, the trade press was near-unanimous that LA multifamily hard costs had hit a ceiling. Lumber had come off its pandemic peak, metals had softened, and the freight index was finally behaving. A reversion to pre-2020 unit economics, so the argument went, was a matter of twelve to eighteen months.

Three years later, on my own book, the per-gross-SF number on a new ground-up is not materially different from where it was in 2023 — it has moved roughly four percent in nominal terms, which, against the cost of capital we are all now paying, is a decline in real terms of essentially zero. The ceiling did not break. The composition underneath it changed.

If your 2026 pro forma is premised on a cost reset, you are underwriting a market that, from where I sit, is not coming.— Field note, March 2026

Where the money went.

The easiest way to see the rotation is to put four years of my own buckets next to each other. The numbers below are a blended average across the five projects, normalized to a gross-square-foot basis and held to the same cost code structure. I am not in the business of publishing line-item data for my jobs, but the directional story is the only thing that matters here.

Read that table honestly and the rotation becomes obvious. The bucket that gets all the press — framing and raw materials — is, in fact, down significantly from its 2022 peak. Everything else is up, and some of it is up a lot. Labor, MEP, carry, and soft costs together now account for a larger share of a new multifamily project in Los Angeles than they have at any point in my thirty-year career.

Labor, and the honesty of a W-2.

Of all the buckets above, labor is the one that deserves the most careful treatment, because it is the bucket where the industry is least honest with itself. What I pay a journeyman electrician, on a W-2, with full burden, is not what the published prevailing-wage table says. It is meaningfully more. It has been meaningfully more for three years. And it is going to stay meaningfully more, because the pipeline behind it — the apprenticeship enrollment, the trade-school throughput, the immigration flows that filled the gap for twenty years — has not kept pace.

The operators I respect most have already adjusted. They have moved their MEP in-house, they have taken on a formal apprenticeship commitment, and they have stopped pretending that a subcontractor quote from 2019 is a useful anchor for a 2026 bid. The rest are still reading the old tables.

A brief digression on self-performance.

Running your own trades is unfashionable. I am aware. The prevailing academic and private-equity view is that a general contractor should be a manager of subcontracts, full stop — that self-performance introduces risk, ties up capital, and distracts from the GC's core competency. I disagree, but I want to be honest about why.

I self-perform because the alternative, in 2026 Los Angeles, is paying a subcontractor to manage a labor problem I am better positioned to manage myself. The arithmetic is simpler than the MBA case studies make it. If I can keep an electrician, a plumber, and an HVAC mechanic productive for forty-eight weeks a year across a portfolio of active jobs, I pay them more than a subcontractor would, and it still costs me less than the subcontractor's bid. That is the whole calculation. I have written about it before.

Soft costs, and the tax of delay.

The last rotation in the table is the one I am angriest about, and the one owners have the least control over: soft costs and permitting. Four years ago, a reasonably clean multifamily permit ran us somewhere between seven and ten months in LADBS. Today, the same package, with arguably cleaner drawings and a more cooperative expeditor, is running fourteen to nineteen.

Every additional month of municipal delay is not just a month of higher carry. It is a month in which the labor pool re-rates, in which material escalators re-set, in which the insurance carrier rewrites your builder's-risk policy. The tax of delay compounds. Our own delay cost, per unit, has roughly tripled against a 2019 baseline — and almost none of that is visible on the hard-cost table.

What owners should do this year.

I am not in the business of writing prescriptions for other people's projects. Every site is different, every capital stack is different, and the honest truth is that most of what matters in a job is unknowable until we walk it. But if I had to pick three dispositions that the composition above should push an owner toward, they would be these.

First, stop anchoring to 2019. The reset the trade press keeps anticipating is not coming, and the operators who will look good in 2028 are the ones who underwrote 2026 costs against 2026 reality, with no implicit reversion baked in.

Second, put a real number on delay. If your pro forma does not have a line item called "municipal risk," or whatever you want to call it, you are leaving out the bucket that is moving fastest. Our jobs budget six to nine months of soft-cost carry above the nominal schedule, and we have rarely found ourselves with too much.

Third, be honest about what the trades are costing you. A W-2 journeyman is what the number actually is. Bid against that, not against what the union hall quoted in 2021. The operators who are solving this problem instead of complaining about it are the ones who will still be in business in 2030.

The ceiling didn't break. It moved. The sooner the rest of the industry says so out loud, the sooner we can all go back to doing the work.

— End of Entry № 042 · Los Angeles, April 9, 2026