Starts Hit 1.487M as Permits Fall to 1.376M: March 12 Release Signals a Pipeline Draw, Not a Boom

The Census Bureau’s March 12, 2026 press release serves a headline sugar high—“starts up”—with a data caveat chaser. January’s total housing starts rose to 1,487,000 SAAR (up 7.2% m/m, 9.5% y/y), yet total permits fell to 1,376,000 SAAR (down 5.4% m/m, 5.8% y/y). The kicker: the starts “increase” isn’t statistically significant at the 90% level. In plain English, the data won’t swear under oath.

Here’s what the release actually shows:

• Total starts at 1,487,000 SAAR (+7.2% m/m, +9.5% y/y), but both changes are flagged as not statistically significant (90% CIs include zero).

• Total permits at 1,376,000 SAAR (−5.4% m/m, −5.8% y/y). The forward pipeline is shrinking.

• Single-family starts slipped to 935,000 SAAR (−2.8% m/m, not statistically significant); single-family permits at 873,000 SAAR (−0.9% m/m).

• 5+ unit starts at 524,000 exceeded 5+ permits (453,000): multi-family is drawing down earlier authorizations.

• Completions increased to 1,527,000 SAAR (+4.8% m/m, not statistically significant), but were −7.5% y/y.

• Backlog (authorized-but-not-started) thinned: seasonally adjusted U.S. total −3.4% m/m; 1-unit flat (0.0% m/m). NSA totals fell from 258.6 (Dec) to 249.3 (Jan); 1-unit was steady (132.7 to 133.0).

• December permits were revised up from 1,448,000 to 1,455,000 SAAR, steepening January’s decline. The release notes average preliminary revisions of about 1.9% or less.

The press release leans on a “starts up” headline, but the math and methodology point the other way: weakening permits, thinning backlogs, and multi-family outperformance that looks more like pipeline run-off than fresh demand.

• Starts beat permits. That’s not expansion; that’s digestion.

• Single-family—the ownership core—lost altitude even as totals rose.

• Completions are still flowing, but the y/y trend is down and the backlog is shrinking.

Quick Scorecard

Starts Beat Permits — But It’s a Mirage

January’s total starts rose to 1.487M SAAR while total permits slid to 1.376M. That divergence is the tell. When starts outpace permits, builders are not laying the groundwork for future growth—they’re finishing what’s already in the queue. The release all but says so: 5+ unit starts at 524,000 eclipsed 5+ permits at 453,000, a classic pipeline drawdown.

Two more red flags:
- The starts increases carry 90% confidence intervals that include zero, flagged by asterisks in the tables. Declaring a solid uptrend from statistically fragile gains risks narrative drift.
- December’s permits were revised up to 1.455M. Using the proper base, January’s −5.4% permit drop is steeper than it first appeared—exactly the opposite of a reacceleration story.

Single-Family’s Soft Center, Multi-Family Does the Heavy Lifting

The core ownership engine—single-family—cooled. Starts fell 2.8% m/m to 935,000 SAAR (again, not statistically significant), while single-family permits eased 0.9% m/m to 873,000. The total headline looks firmer only because multi-family kept punching above its weight, but even that strength looks borrowed from prior authorizations rather than freshly minted permits.

Under the hood:
- Single-family completions slipped 1.0% m/m to 970,000 (not statistically significant), consistent with a moderating pipeline.
- The sector mix matters. A total starts figure buoyed by 5+ activity can mask weakening demand in the ownership segment that drives land spend, spec inventory, and broad construction employment.

Completions Are Up—Yesterday’s News

Total completions climbed to 1.527M SAAR (+4.8% m/m, not statistically significant), but they’re −7.5% y/y and out of sync with permits. That’s the backward-looking face of housing: deliveries still look solid because they reflect decisions made quarters ago. Today’s permits say tomorrow’s output slows.

Backlog metrics confirm the turn:
- Seasonally adjusted authorized-but-not-started totals declined 3.4% m/m; 1-unit was flat.
- NSA totals fell from 258.6 (Dec) to 249.3 (Jan); 1-unit barely moved (132.7 to 133.0).
- Regionally (NSA), permits fell from December to January across the Northeast (20.4→20.1), Midwest (21.9→21.5), South (154.0→148.2), and West (62.3→59.5). The West’s total permits saw a notable monthly drop (31.4→21.3, NSA). One bright spot: South 1‑unit permits ticked up (36.5→38.1) even as the U.S. total 1‑unit slipped (63.6→61.9).

Put together, this is how a cycle cools: permits down, backlog thinner, completions still okay—until they’re not.

Revisions, Seasonality, and the Winter Fog

Winter data are slippery. The release lifts December permits to 1.455M from 1.448M, and it reminds us that preliminary SAAR estimates are typically revised by around 1.9% or less. That’s material enough to bend a month-to-month story line. Context helps: January 2026 starts at 1.487M rhyme with February 2025’s 1.490M level, but the permit backdrop is weaker now than in January 2025 (1.376M vs 1.460M). Similar headlines, softer foundations.

It’s also why the asterisk matters. When the 90% confidence interval includes zero, prudence beats punditry. The “up” in starts and completions may be noise rather than a new trend.

What This Means for Markets

Equities: Builders, Materials, and REITs
- Homebuilders: The mix shift toward multi-family and the dip in single-family permits (873k) argue for caution on broad beta. Favor operators with high optioned land, flexible spec strategies, and outsize exposure to the South, where 1-unit permitting (NSA) bucked the decline. Be selective; volume growth looks scarce without a rate tailwind.
- Building products and aggregates: Near-term deliveries (1.527M completions) support 1H order books, but the permit/backlog slide warns of a 2H fade. Prioritize pricing power and repair/remodel exposure over new-construction dependence.
- Apartment REITs: Supply pressure from the multi-family wave is still landing, but 5+ permits (453k) lag 5+ starts (524k). As pipeline runoff accelerates, supply relief could firm rent growth into late 2026—especially in markets outside the most supply-heavy Sun Belt submarkets.

Rates, Credit, and Housing Finance
- MBS/Treasuries: Softer forward supply and a weakening single-family engine tilt modestly disinflationary on shelter in 2H—supportive at the margin for duration, contingent on labor and services inflation. Watch for rent growth stabilization rather than a re-acceleration.
- Construction lending and multifamily credit: The 5+ imbalance (starts > permits) implies peak deliveries now, thinner later. Underwriting should refocus on lease-up velocity and concessions risk through mid-2026; tightening spreads for projects with superior absorption and fixed-rate debt.
- RMBS/ABS: With single-family starts at 935k and permits slipping, expect subdued origination volumes; prepayment risk stays muted absent a sharp rate rally.

Macro Watchlist
- Mortgage rates and absorption: Without a material rate reset, single-family’s −0.9% m/m permit dip may persist. Absorption metrics will determine whether builders keep leaning on incentives or step back on starts.
- Regional pulse: The West’s NSA permit slump (31.4→21.3) warrants scrutiny for local builders, land values, and municipal fee revenues. The South’s resilient 1-unit permits are the outlier to track for relative performance.
- Revisions and significance flags: Treat winter SAAR prints with humility; the 1.9% revision norm and the “not statistically significant” stamps argue for trend confirmation before repositioning.

The Investor Takeaway

• Position for a softer 2H supply tape. The combination of permits down (1.376M), a thinning backlog (−3.4% SA m/m), and completions still catching up sets up a later-year deceleration in deliveries.

• In builders, prefer balance-sheet strength and land-light optionality over volume chasers. Geographic bias to the South remains sensible given relative permit resilience.

• For apartments, the 5+ pipeline drawdown (starts 524k vs permits 453k) tilts 2026–27 rent risk upward from depressed levels—favor quality operators in submarkets past peak deliveries.

• For materials, lean into pricing power and R&R exposure; fade pure-play new-build cyclicality into the back half.

• In rates and credit, the data skew away from a construction-led reacceleration. Duration risk looks manageable; in credit, emphasize projects with clear absorption and less refi exposure.

Headlines will celebrate higher starts. The March 12 data say something else: the market is eating through yesterday’s permits while tomorrow’s pipeline shrinks. That’s not a boom. It’s a baton pass—from backlog to scarcity—and investors should reposition before the handoff completes.