Housing Starts vs Building Permits

January 2026 housing starts: 1.487 million seasonally adjusted annual rate, +7.2 percent month-over-month, +9.5 percent year-over-year (vs January 2025 1.358M). January 2026 building permits: 1.38 million annualized, -5.4 percent month-over-month, wiping out December gain.

The combined January 2026 reading: starts > permits (1.487M > 1.38M, ~107K gap). Builders working through existing permit backlog without replenishing pipeline. Pattern suggests Q2 2026 starts likely to moderate as permit drawdown effect propagates.

February and March 2026 data delayed: originally scheduled March 17 (February release) and April 17 (March release) postponed to April 29, 2026 single combined release.

The combined April 2026 reading: housing market in late-cycle resilience phase. Strong starts despite mortgage rates 5.98-6.22 percent (Q1 2026 average 6.0-6.3 percent). Permit weakness suggests builder caution about pipeline depth. The starts-permits gap is narrowing toward 1.40-1.45M sustainable annualized rate.

Building permits lead housing starts by 3-6 months typically. Process: builder pulls permit (legal authorization to begin construction); construction starts within 3-6 months (depends on builder schedule, weather, regional regulations); construction completes 6-12 months after starts.

The practical implication: monitoring permits provides early warning for starts trajectory. Permit decline 6 months ago = starts decline ahead. Permit recovery 6 months ago = starts recovery ahead.

Empirical lead-lag. 2007-2008 housing crisis: permits peaked April 2006 at 2.26M annualized; starts peaked April 2006 at 2.27M. Permits fell to 1.32M by January 2008 (-42 percent); starts fell to 1.07M by January 2008 (-53 percent). Permits led starts decline by 1-2 months.

2020 COVID flash crash: permits fell from 1.42M (February 2020) to 1.04M (April 2020); starts fell from 1.62M (February 2020) to 0.93M (April 2020). Permits and starts moved synchronously in pandemic shock.

2022-2023 hiking cycle: permits fell from 1.97M (December 2021) to 1.35M (January 2023) -32 percent. Starts fell from 1.79M (April 2022) to 1.30M (January 2023) -27 percent. Permits led starts decline by 4 months.

2024-2026 stabilization: permits rose from 1.44M (December 2023) to 1.55M peak then fell to 1.38M (January 2026). Starts ranged 1.35-1.50M same period.

Housing starts and permits split into single-family and multi-family categories. Single-family is approximately 70 percent of total in 2024-2026 era; multi-family approximately 30 percent.

Single-family dynamics: dominated by mortgage rate sensitivity (affordability). 30-year fixed mortgage rate 5.98-6.22 percent (Q1 2026) versus 3.0 percent (2021). Affordability stress has compressed single-family starts to ~0.95-1.05M annualized vs 2021 peak ~1.15M.

Multi-family dynamics: driven by rental demand, construction loan availability, regional housing shortage. Multi-family starts have been more volatile. 2022-2023 multi-family starts spiked above 600K annualized as rental demand surged. 2024-2025 corrected to 350-400K as rental absorption slowed and construction lending tightened.

The practical implication: single-family-heavy starts decline reflects mortgage rate stress; multi-family-heavy starts decline reflects rental demand cycle. Both contributed to 2022-2023 starts compression.

April 2026 starts mix: estimated single-family ~1.0M, multi-family ~485K (1.487M total). Single-family stable; multi-family rebounded modestly from 2024 lows reflecting rental demand recovery in Sun Belt + Mountain West markets.

Starts and permits typically move together but can diverge during specific regimes.

Divergence regimes. Builders building from inventory permits: starts > permits (current phase, January 2026 starts 1.487M > permits 1.38M). Suggests builders confident enough to start but not pulling new permits. Forward-looking: starts will moderate as inventory permits exhaust.

Builders accumulating permits without starting: permits > starts (typical of weather-impacted Q4-Q1 in cold regions, then Q2-Q3 starts surge). Forward-looking: starts surge ahead.

Regional effects: hurricane regions show permits fall after major storms (regulatory chaos), then surge as rebuilding begins. Cold regions show seasonal permits-starts cycle.

Long-run starts/permits ratio: 0.97-1.05 (very close to 1.0). Permits slightly higher than starts on average due to abandoned permits (builders don't always start construction).

April 2026 ratio: 1.487M / 1.38M = 1.078 (above long-run range, indicating pipeline drawdown). Forward-looking: starts likely to moderate to ~1.40M annualized in Q2-Q3 2026 unless permits accelerate.

Three macro cycle examples of starts-vs-permits dynamics.

2005-2008 housing crisis: permits peaked 2.26M (April 2006); starts peaked 2.27M (April 2006). Both fell ~50 percent peak-to-trough. Permit lead time approximately 1-2 months. Starts/permits ratio averaged 0.99 throughout.

2020 COVID flash crash: permits fell from 1.42M (February 2020) to 1.04M (April 2020). Starts fell from 1.62M to 0.93M. Both recovered quickly: permits 1.40M by August 2020, starts 1.42M by August 2020.

2022-2023 hiking cycle: permits peaked 1.97M (December 2021); starts peaked 1.79M (April 2022). Both fell substantially. Permits to 1.35M (January 2023) -32 percent; starts to 1.30M (January 2023) -27 percent. Starts/permits ratio fell to 0.96 (permits ahead of starts decline).

2024-2026 stabilization: permits ranged 1.38-1.55M; starts ranged 1.35-1.49M. Mortgage rate stabilization at 6-7 percent supported residual housing demand.

The pattern: starts and permits move together with permits leading by 3-6 months at major cycle turns. Long-run starts/permits ratio 0.97-1.05.

Stress history shows specific starts-vs-permits patterns.

2007-09 housing crisis: starts fell from 2.27M (April 2006) to 0.48M (April 2009) -79 percent. Permits fell from 2.26M (April 2006) to 0.51M (March 2009) -77 percent. Permits led by 1-2 months consistently throughout decline.

2020 COVID flash crash: starts fell from 1.62M (February 2020) to 0.93M (April 2020) -43 percent. Permits fell from 1.42M to 1.04M -27 percent. Starts fell more than permits as construction shutdowns affected ongoing projects more than new permits. Recovery: both above pre-COVID levels by August 2020.

2022-2023 hiking cycle: starts fell from 1.79M (April 2022) to 1.30M (January 2023) -27 percent. Permits fell from 1.97M (December 2021) to 1.35M (January 2023) -32 percent. Permits led starts decline by 4 months.

2026 mortgage rate stabilization: starts +7.2 percent in January 2026 to 1.487M. Permits -5.4 percent to 1.38M. Divergence reflects builders working through inventory permits; pipeline drawdown signal.

The pattern: starts-permits relationship breaks during shock-driven shutdowns (2020 COVID). Otherwise permits lead starts by 1-4 months consistently across cycles.

Housing starts and permits are not directly tradable but drive significant moves in homebuilders (XHB, ITB) and building materials (Vulcan, Martin Marietta). Census Bureau releases starts and permits jointly monthly (typically third Tuesday or Wednesday of month).

Released data revisions: housing starts and permits frequently revised up or down by 5-15 percent in subsequent month. Revisions reflect updated regional data and sample completeness. The volatility of revisions makes single-month readings noisy. Three-month moving averages provide cleaner signal.

For positioning around starts/permits releases. Strong release (above consensus): XHB rallies 1-3 percent typically; building materials (VMC, MLM) outperform; mortgage REITs (NLY, AGNC) marginally affected. Weak release (below consensus): XHB falls 1-3 percent; building materials underperform; mortgage REITs marginally negative.

The practical implication: starts/permits data drives short-term moves in housing-related stocks but does not strongly affect broader equity indices except during major housing cycle inflection points (2007-2008, 2020 COVID).

For longer-term positioning: monitor starts/permits trend via 3-month moving averages. Sustained decline below 1.30M signals housing recession; sustained increase above 1.50M signals housing acceleration.

For macro allocators, starts-vs-permits provides housing cycle classification.

Starts > permits + both above 1.50M: late-cycle housing strength with pipeline drawdown. Forward-looking: starts moderation ahead.

Starts < permits + both above 1.50M: pipeline depth strong. Forward-looking: starts acceleration ahead.

Starts < permits + both below 1.30M: weak pipeline, recession-imminent. Forward-looking: starts decline ahead.

Starts > permits + both below 1.30M: pipeline drawdown in weakness. Forward-looking: starts trough nearing.

April 2026 setup: starts 1.487M > permits 1.38M, both in 1.30-1.50M range (mid-trend). Configuration suggests late-cycle pipeline drawdown but no recession signal. Starts likely to moderate to ~1.40M absent permit acceleration.

For housing-related positioning: long XHB benefits from sustained starts above 1.40M; long building materials (VMC, MLM) benefits from infrastructure capex + housing stability; long mortgage REITs (NLY, AGNC) benefits from Fed cuts compressing rates. Short positions warranted when permits decline sustained 3+ months and approach 1.30M annualized rate.

Starts-vs-permits captures construction pipeline depth. Compared to other housing indicators.

Vs new home sales: new home sales (Census Bureau) measures contracts signed for newly built homes. Lagging indicator (3-6 months from start to sale). Starts/permits leads new home sales.

Vs existing home sales: existing home sales (NAR) measures closed transactions for existing homes. Different market segment (resale vs new build). Less direct relationship to starts/permits.

Vs Case-Shiller home price index: home price appreciation index. 3-month moving average lagging indicator. Starts/permits leads home prices by 6-12 months.

Vs NAHB builder confidence: NAHB Housing Market Index measures builder sentiment. Leading indicator (1-3 months ahead of permits). NAHB current April 2026 reading not yet specified in research but typically 35-50 range during current era.

Vs XHB stock performance: XHB price action leads starts/permits by 1-3 months as market discounts forward fundamentals. April 2026: XHB $109.44 near 52-week highs.

For allocator monitoring, starts-vs-permits is foundational housing supply indicator. Pair complements XHB price (forward-looking), NAHB confidence (leading), new home sales (lagging), Case-Shiller (price impact) for comprehensive housing cycle read.

January 2026 starts 1.487M (+7.2% MoM, +9.5% YoY); January 2026 permits 1.38M (-5.4% MoM). February and March 2026 data delayed to April 29, 2026 combined release.

Forward-looking through 2026: the April 29 release will provide first read on Q1 2026 housing trajectory. Three scenarios.

Base case: starts moderate to 1.40-1.45M annualized in February-March; permits stable around 1.40M. Configuration suggests housing stable at sub-trend levels.

Upside scenario: starts above 1.50M; permits above 1.45M. Mortgage rate decline below 6 percent (February 2026 reading) catalyzed both. Configuration suggests housing recovery accelerating.

Downside scenario: starts below 1.30M; permits below 1.30M. Mortgage rate persistence above 6.5 percent + builder margin compression hit both. Configuration suggests housing recession emerging.

Key watches: 30-year mortgage rate direction (current 5.98-6.22 percent); builder confidence (NAHB monthly); homebuilder earnings (Lennar Q2 2026 results); construction lending availability for multi-family.

Expected starts/permits range 1.35-1.55M absent major catalyst. Pipeline drawdown pattern (current January 2026) likely to continue through Q2 2026 unless permit acceleration emerges.