Macro ReviewHow to read
Thematic · Housing supply 2027–28

Shelter slows through 2027 — 3.9% now; the risk is 2028

Builders are finishing a large backlog of new homes now, which holds shelter inflation down; the question is what happens when that pipeline runs dry.Does the shelter disinflation in the pipeline run out in 2028 — and does a hike make that worse?

Said 13 Sept 2026 · revised 15 Sept 2026Data to Oct 2026 (shelter) · new tenant rents to Q3 2025, paused · updated 17 Nov 2026
Shelter CPI, 3-month annualised+3.9%Oct 2026 · 12-month 3.4% · 6-month 3.1% · from Shelter and rents · trigger 3.5% · crossed
CPI rent, 6-month annualised+2.8%Oct 2026 · 3-month 2.5% · 12-month 2.9%
New tenant rent, 4 quarters+1.2%Q3 2025 · 95% interval −4.0 to 6.5% · last edition · paused since Apr 2026
Rule: CPI rent y/y, Q3 2026+2.5%from new tenant rents 4 quarters earlier · band 1.1 to 3.9%a rule, not a release
Completions, 5+ unit buildings339Sep 2026 · thousands, annual rate · from the Construction pipeline page
of new homes8.4Sep 2026 · 2022 analogue 9.4 · from the New-home supply page · trigger below 8.0 by Q1 2027 · not crossed

Shelter slows through 2027; the risk is 2028

Does the shelter disinflation in the pipeline run out in 2028 — and does a hike make that worse?
  • Shelter slows through 2027; the risk is 2028, when new apartment supply stops arriving: tested (said 13 Sep 2026).→ Exhibit 2
  • Shelter CPI ran 3.9% over the three months to Oct 2026, against a 3.5% trigger.→ Exhibit 5
  • New tenant rents rose 1.2% over the four quarters to Q3 2025, the last edition before BLS paused the index in April 2026.→ Exhibit 1
  • Starts already made imply 380k completions over the next year, against 413k in the last.→ Exhibit 9
  • New-home is 8.4 (Sep 2026) against the ’s 8.0 test for Q1 2027.→ Exhibit 12
What would change this view
What would change this viewTriggerTodayDecided by
Shelter re-accelerates nowabove 3.5% at a three-month annualised rate3.90% (Oct 2026); on 11 Sep 3.22% (Sep 2026)crossedConsumer Price Index, 10 Dec 2026
The housing objection to a sequence fallsnew-home months' supply below 8.0 by Q1 20278.4 months (Sep 2026)not crossedNew Residential Sales, 25 Nov 2026
The reasoning behind each point
  • Verdict
    Shelter slows through 2027; the risk is 2028, when new apartment supply stops arriving: tested (said 13 Sep 2026).

    As written on 13 Sep 2026: “A hike does not raise shelter CPI in 2027 — the pipeline does the opposite — but it raises the 2028 problem, and it moves the marginal household from owning to renting in a market where multifamily deliveries are about to halve.” The verdict is “tested” because a trigger of the falsifier table has crossed; the scorecard sets what the log said against what has happened since.

  • Shelter
    Shelter CPI ran 3.9% annualised over the three months to Oct 2026, against a 3.5% trigger.

    Over twelve months shelter ran 3.4%. Rent 2.5% over three months and 2.8% over six (2.9% on the year, unadjusted); owners' equivalent rent 3.5% and 3.4% (3.4%). The three-month pace is at or above the six-month: momentum is not falling.

  • Early warning
    New tenant rents rose 1.2% over the four quarters to Q3 2025, the last edition before BLS paused the index in April 2026.

    The 95% interval runs from −4.0 to 6.5%. The rent rule maps new tenant rents to CPI rent four quarters later (correlation 0.89) and puts CPI rent at 2.5% y/y in Q3 2026 (band 1.1 to 3.9%). CPI rent was 2.8% over four quarters to Q3 2026. No index covers new leases after Q3 2025; the private indices the September report quotes disagree (Zillow +2.5%, Apartment List −0.8%).

  • Pipeline
    Starts already made imply 380k multifamily completions over the next year, against 413k in the last.

    Completions in 5+ unit buildings ran at 339k a year in Sep 2026. Real residential investment −2.5% over four quarters to Q3 2026 (GDP & GDI page). The September report’s delivery forecasts — 444,343 units in 2027 on Yardi Matrix, 290–339k on CoStar/MMG — are quoted, not tracked. The projection is the Construction pipeline page's distributed-lag model, copied here: it holds starts after the latest month at their three-month average and is model-conditional.

  • Months' supply
    New-home months' supply is 8.4 (Sep 2026) against the September report’s 8.0 test for Q1 2027.

    Rising into the decision in 1973 and 2006, when the tightening was reversed; falling through 2022–23, when housing absorbed it. The September report’s falsifier for the housing objection is new-home months' supply below 8.0 by Q1 2027.

Next: Financing & affordability — mortgage rates and payments

Source: BLS CPI and research rent indexes; Census; BEA; the September 2026 report · data to the dates shownthemes.housing_supply-read

Exhibit 2 · The September report's paths and tests against what has happened since

Each named test and path set: what the dated record said on 13 Sep 2026, what has happened since, and the verdict
WhatSaidHappenedVerdict
PathsWhere shelter inflation goes, 2027–28Said · 13 Sept 20263 paths named, no odds: Down through H1 2027 — the pipeline · Re-acceleration as the delivery trough arrives — H2 2027 into 2028 · Re-acceleration now — the falsifier pathHappenednot yet decided — the September report’s dated outlook for CPI shelter and the path that would falsify itVerdictopen
PathsWhich tightening episode housing resembles at the decisionSaid · 13 Sept 20262 paths named, no odds: 1973 and 2006 — housing already turning; the tightening was reversed · 2022 — housing absorbed the tighteningHappenednot yet decided — the September report’s discriminator across eight supply-shock episodesVerdictopen
PathsOne move or a sequence into the 2028 cliffSaid · 13 Sept 20262 paths named, no odds: One move, framed as a recalibration · A sequenceHappenednot yet decided — what the policy path does to shelter by 2028Verdictopen
TestShelter re-accelerates nowSaid · 13 Sept 2026above 3.5% at a three-month annualised rateHappened · Oct 20263.90% (Oct 2026); on 11 Sep 3.22% (Sep 2026)Verdictcrossed
TestThe housing objection to a sequence fallsSaid · 15 Sept 2026new-home months' supply below 8.0 by Q1 2027Happened · Sep 20268.4 months (Sep 2026)Verdictnot crossed
TestMarket rents turn upSaid · 13 Sept 2026Apartment List rents above +2% y/y by DecemberHappenedsubscription data — quoted from the September report, not trackedVerdictnot tracked here
TestExisting-home supply tightensSaid · 13 Sept 2026NAR months' supply below 4.0Happenedsubscription data — quoted from the September report, not trackedVerdictnot tracked here

The view is tested: the word follows a rule set out in the Appendix, from the log and the tests above, never by hand.

About this exhibit
A scorecard of the path sets and the tests named in the September report, each with what was said, what has happened and a verdict
What was said on 13 Sept 2026, and what has happened since
The three sets of paths and the four tests of the view: what the dated log said, what the data show, and the verdict
Source: the theme's dated log (said 13 Sept 2026); BLS; Census · data to Oct 2026themes.housing_supply-22

Exhibit 3 · The timeline of new-lease rents, apartment supply and the hike

When new-lease rents reach the CPI, apartment deliveries bottom out, the hike's push into renting peaks, and each named test is decided, through 2027–28
2025202620272028Rentsnew-lease rents flat…reach CPI rent a year later: slowingrule: 2.5% (Q3 2026)Supplyapartmentsdeliveries high2027: the delivery lowrents lose that supplyThe hike17 Sep 2026buyers pushed into rentingpeaks 12–18 months onTestsOct 2026: shelter above 3.5%? ▲Q1 2027: new-home supply below 8.0 months? ▲
Row 1: new-lease rents lead CPI rent by about 4 quarters, so the CPI keeps slowing through 2027 (the rule puts CPI rent at 2.5% in Q3 2026). Row 2: the apartments started in the boom are delivered, the flow reaches a low in 2027 and stops holding rents down in 2028. Row 3: the hike's push from owning to renting peaks twelve to eighteen months on — in the same year.
About this exhibit
A timeline diagram in four rows: new-lease rents and when they reach CPI rent; apartment deliveries and their low; the hike and when its push into renting peaks; the dates the tracked tests are decided
Two lags meet in 2028: late rents and vanishing supply
The calendar of the argument, 2025 to 2028: what leads what, and when the tests of the view are decided
Source: BLS research rent indexes and CPI; Census new residential construction; the September 2026 report · data to Oct 2026themes.housing_supply-23

Exhibit 4 · The view, and the claim as written

The author's claim on 13 Sept 2026, kept as written; a figure revised since is marked where it stands

Shelter slows through 2027; the risk is 2028, when new apartment supply stops arriving — tested

As written on 13 Sept 2026

A hike does not raise shelter CPI in 2027 — the pipeline does the opposite — but it raises the 2028 problem, and it moves the marginal household from owning to renting in a market where multifamily deliveries are about to halve.

Multifamily (5+ unit) completions ran at 329,000 a year in July 2026 (revised: 357k (Census, 20 Oct 2026)), 25.6% lower than a year earlier; Yardi Matrix puts 2027 deliveries at 444,343 units, the cycle low, so the supply holding rents down in 2026–27 stops arriving in 2028. CPI shelter ran 2.1% annualised over the three months to August while Zillow's rent index re-accelerated to +2.5%. The BLS New Tenant Rent Index — the early-warning series — has been paused since April 2026. New-home months' supply was 9.6 in July, level with November 1973 and above June 2022's 9.4.

Restated in the 15 September edition as a path argument: a single hike lands its shelter effect in the same year as the delivery cliff; a sequence lands four of them there.

StatusactiveSaid13 Sept 2026Revised15 Sept 2026

Argued in the : 6 dated documents, listed with their sections under Sources in the Appendix.

Source: the September 2026 report and its research notes (6 documents, listed in the Appendix)themes.housing_supply-1

Rents are still slowing; deliveries are falling

What do shelter, new-lease rents and the construction pipeline say today?

Exhibit 5 · Shelter CPI's three-month pace against twelve-month shelter, rent and owners' equivalent rent

% change, monthly, to Oct 2026 · shelter annualised over three months, and shelter, rent and OER over twelve months · dashed line: the 3.5% test
About this exhibit
Monthly percent change: shelter CPI annualised over three months, and shelter, rent and owners' equivalent rent over twelve months, with a dashed line at the three-month test
Shelter CPI: the three-month pace against the twelve-month rate
% change, annualised or on a year earlier, Oct 2021 to Oct 2026
Source: U.S. Bureau of Labor Statistics, CPI shelter indexes. From Shelter and rents (data to Oct 2026)themes.housing_supply-2

Exhibit 6 · Inflation by horizon in shelter, rent and owners' equivalent rent

% change in Oct 2026: three-month and six-month rates annualised (seasonally adjusted), and over twelve months (unadjusted)
Index3 months6 months12 months
Shelter3.9%3.1%3.38%
Rent of primary residence2.5%2.8%2.90%
Owners' equivalent rent3.5%3.4%3.43%

The report's test is shelter above 3.5% on the three-month rate. A three-month pace below the six-month means the pipeline disinflation is still coming through.

About this exhibit
A table of shelter, rent of primary residence and owners' equivalent rent: percent change annualised over three and six months and over twelve months, in the latest month
Shelter momentum, Oct 2026
Annualised over three and six months (seasonally adjusted) and over twelve months (unadjusted), as the shelter figures publish them
Source: U.S. Bureau of Labor Statistics, CPI shelter indexes. From Shelter and rents (data to Oct 2026)themes.housing_supply-3

Exhibit 1 · Rent growth for new tenants against all tenants and CPI rents

% change on four quarters earlier, quarterly, Q1 2015 to Q3 2026 · dashed line: this site's rule, CPI rent projected from new-tenant rents 4 quarters earlier
About
About this exhibit
Four-quarter change in five rent series: BLS new-tenant rent, all-tenant rent, CPI rent, CPI owners' equivalent rent, and this site's rule (CPI rent projected from new-tenant rents some quarters earlier, dashed)
New tenant rents lead CPI rent — and stopped at Q3 2025
% change on four quarters earlier, Q1 2015 to Q3 2026
Source: U.S. Bureau of Labor Statistics research indexes R-CPI-NTR and R-CPI-ATR; rent rule calculated on this site. From Shelter and rents (data to Jul 2026)themes.housing_supply-4

Exhibit 7 · The last new-tenant rent reading, CPI rent and the rule's projection

% change on four quarters earlier: new-tenant index to Q3 2025 (BLS has paused it), CPI rent to Q3 2026, the rule's figure for Q3 2026
New tenant rent, 4 quarters to Q3 2025+1.2%
95% interval-4.0 to 6.5%
CPI rent, 4 quarters to Q3 20262.76%
Rule: CPI rent y/y, Q3 2026 (this site's rule)2.54%
Rule: correlation at a 4-quarter lead · residual σ0.89 · 0.71pp

No official index covers new leases after Q3 2025. The private indices the report quotes disagree in sign; their readings are in the quoted row “Market rents” below (subscription data, quoted, not tracked).

About this exhibit
A table: new-tenant rent's last four-quarter change with its 95% interval, CPI rent over four quarters, the rule's CPI rent figure for the next quarters, and the rule's correlation and residual error
The early-warning index is dark
Last edition Q3 2025; BLS paused publication in April 2026
Source: U.S. Bureau of Labor Statistics research indexes R-CPI-NTR and R-CPI-ATR; rent rule calculated on this site. From Shelter and rents (data to Jul 2026)themes.housing_supply-5

Exhibit 8 · The multifamily pipeline: starts, completions and units under construction

thousands of units; starts and completions at an annual rate, Sep 2016 to Sep 2026
Source: U.S. Census Bureau and HUD, New Residential Construction. From Construction pipeline (data to Sep 2026)themes.housing_supply-6

Exhibit 9 · Multifamily completions and the completions implied by starts already made

Multifamily completions, thousands at an annual rate, to Sep 2026 · dashed: the Construction pipeline page's projection and band, starts held at their three-month average
About this exhibit
Completions of buildings with five or more units, thousands at an annual rate, monthly, and the Construction pipeline page's model projection from starts already made, with its band
Completions implied by starts already made — model-conditional
Multifamily completions, thousands at an annual rate, to Sep 2026, and the Construction pipeline page's projection with its 90% band
Source: Census new residential construction; the Construction pipeline page's distributed-lag model · data to Sep 2026themes.housing_supply-24

Exhibit 10 · Months to finish the units under construction at the current completion rate

Months, monthly, to Sep 2026: units under construction over completions at an annual rate, all buildings and multifamily
About this exhibit
Units under construction divided by completions at an annual rate, in months, for all buildings and for buildings with five or more units
How long the pipeline takes to clear
months of completions at the current rate, Sep 2016 to Sep 2026
Source: U.S. Census Bureau and HUD, New Residential Construction. From Construction pipeline (data to Sep 2026)themes.housing_supply-7

Exhibit 12 · New-home months' supply against the September report's test

Months of sales at the current pace, monthly, Jan 2015 to Sep 2026 · dashed line: the 8.0 test, by Q1 2027
About this exhibit
New-home months' supply, monthly, with a dashed line at the September report's test
Months' supply at the decision
months, Jan 2015 to Sep 2026
Source: U.S. Census Bureau and HUD, New Residential Sales. From New-home supply (data to Sep 2026)themes.housing_supply-9

Exhibit 14 · Real residential investment growth and residential investment price inflation

% change on four quarters earlier, and real growth quarter annualised (dashed), quarterly, to Q3 2026
About this exhibit
Real residential investment, percent change on four quarters and quarter annualised (dashed), and residential investment prices on four quarters
Residential investment — the building that becomes 2028's supply
% change, Q3 2021 to Q3 2026
Source: U.S. Bureau of Economic Analysis, NIPA Table 1.1.6 and 1.1.5. From GDP & GDI (data to Jul 2026)themes.housing_supply-11

Exhibit 15 · Revisions since the September report

The report's figures against the housing pages' current estimates, same month
SeriesMonthReportNowRevision
Housing starts, TotalJul 20261,239.01,369.0+130.0
Completions, TotalJul 20261,212.01,283.0+71.0
Under construction, TotalJul 20261,262.01,267.0+5.0
Completions, 5 units or moreJul 2026329.0357.0+28.0
Housing starts, 5 units or moreJul 2026421.0484.0+63.0
Under construction, 5 units or moreJul 2026666.0667.0+1.0
Months' supply, publishedJan 20269.99.9±0.0
Months' supply, publishedMar 20268.78.7±0.0
Months' supply, publishedMay 20269.39.2−0.1
Months' supply, publishedJun 20268.58.5±0.0
Months' supply, publishedJul 20269.68.9−0.7
CPI shelter, three months annualisedAug 20262.102.10±0.00
CPI rent, three months annualisedAug 20262.32.3±0.0
CPI rent, six months annualisedAug 20263.43.4±0.0
Owners' equivalent rent, three months annualisedAug 20262.82.8±0.0
Owners' equivalent rent, six months annualisedAug 20263.73.7±0.0
CPI shelter, six months annualisedAug 20263.53.5±0.0
CPI rent, twelve monthsAug 20262.72.8+0.1
Owners' equivalent rent, twelve monthsAug 20263.13.1±0.0
Source: the September 2026 report (13 Sep); Census, via the Construction pipeline and New-home supply pages; BLS, via the shelter series this page carriesthemes.housing_supply-12

Exhibit 16 · Housing figures the September report quoted, by source and date

Named sources' housing figures as the September report quoted them on 13 Sep 2026: Census, delivery forecasts, private rents, builders and the policy channel
WhatAs quotedSource
Multifamily pipeline (Census, July 2026)published data5+ unit completions 329,000 at an annual rate, −25.6% y/y (Mar 455k → Apr 540k → May 418k → Jun 386k → Jul 329k); 5+ units under construction 666,000 (−4.4%); 5+ starts 421,000 (−7.1%). Released 18 Aug 2026.research note: Hike or hold markets (13 Sep 2026) · §3.6 (Census/HUD New Residential Construction; FRED COMPU5MUSA) · 13 Sept 2026
Total pipeline (Census, July 2026)published dataHousing starts 1,239,000 (−13.5% y/y); completions 1,212,000 (−16.8%); units under construction 1,262,000.the September 2026 report · §1.6c key points · 13 Sept 2026
Delivery forecastssubscription dataYardi Matrix Q3 supply forecast (reported 25 Aug): 2027 multifamily deliveries 444,343 units, the cycle low; 2028 modestly above 2027; a return to 2024–25 levels 'highly unlikely'. CoStar/MMG: roughly 695k (2024) → 531k (2025) → 382–407k (2026) → 290–339k (2027). The September report prints both without reconciling them.research note: Hike or hold markets (13 Sep 2026) · §3.6; REPORT §1.6c key points · 13 Sept 2026
Market rentssubscription dataZillow ZORI August $1,948, +2.5% y/y, 'reaccelerating, up from 2.3% last month and 2% a year ago', concessions on 39.2% of listings; Apartment List August −0.8% y/y, vacancy 7.1%; RealPage Q2 effective asking rents −0.2%, occupancy 95.5%; CBRE Q2 vacancy 4.3%.research note: Inflation sources outlook (13 Sep 2026) · §3 Shelter and housing services · 13 Sept 2026
Absorptionsubscription dataRealPage: Q2 completions 77,700 against a peak of about 160,000 in Q3 2024; completions 340,200 in the year to Q2, below the decade norm. The September report: Q2 absorption exceeded deliveries by more than 45,000 units.research note: Inflation sources outlook (13 Sep 2026) · §3; REPORT §1.6c key points · 13 Sept 2026
The early-warning indexpublished dataBLS: 'Due to a lapse in appropriations resulting in uncollected CPI Housing Survey data for October 2025 and competing priorities, BLS paused publication of the R-CPI-NTR and R-CPI-ATR data in April 2026.' Latest edition 2025 Q3; pause announced 24 Apr 2026.research note: Hike or hold markets (13 Sep 2026) · §3.6 · 13 Sept 2026
New-home months' supplypublished data9.9 (Jan 2026) → 8.7 (Mar) → 9.3 (May) → 8.5 (Jun) → 9.6 (Jul); at earlier decisions 9.4 (Oct 1973) and 9.6 (Nov 1973), 6.3 (Jun 2006), 8.2 (Aug 1990), 9.4 (Jun 2022, falling to 7.5 by Jun 2023).research note: Hike or hold case (13 Sep 2026) · §4.3 (FRED MSACSR, T1) · 13 Sept 2026
Existing homes and builderssubscription dataNAR August: sales 3.98m SAAR (−1.2% y/y), inventory 1.62m (+5.9%), months' supply 4.9, the highest in over a decade. Redfin: 57.9% more sellers than buyers. NAHB sentiment 35; 35% of builders cutting prices (average 6%), 63% using incentives; Lennar incentives 12.9% of price on a 15.6% gross margin.the September 2026 report · §1.6c key points · 13 Sept 2026
The policy channelpublished dataDias and Duarte (Fed IFDP 1248): a 25bp contractionary shock raises rents by roughly 0.4–0.5%, peaking at six to twelve months, through the tenure shift. FHFA WP 24-03 (lock-in): 18.1% lower probability of sale per point of mortgage-rate gap, 1.33m sales prevented. 30-year mortgage 6.76%; Fed MPR: 'the majority of outstanding mortgages still have interest rates below 4 percent.'research note: Hike or hold case (13 Sep 2026) · §0.3; REPORT §1.6c · 13 Sept 2026
About this exhibit
A table of figures quoted from the September report's documents, each with source and date: the Census pipeline as printed, delivery forecasts, private rent indices, absorption, the BLS pause notice, months' supply history, existing homes and builders, and the policy channel
Pipeline and market rents beyond these pages, as the report quoted them
Census figures as the report printed them, delivery forecasts, private rent indices, builders and the policy channel; quoted with date and source
Source: the September 2026 report's documents, quoted with dates; not a data feedthemes.housing_supply-13

Market rents

How far ahead do new-lease rents lead?

Exhibit 17 · Correlation of CPI rents with earlier new-tenant rents, by lead

Correlation of four-quarter changes, a ratio: CPI rent from Q1 2006 (and to Q4 2019) and OER, with new-tenant rents 0 to 8 quarters earlier
About
About this exhibit
Bars and a line: correlation of the four-quarter change in CPI rent (full sample, and the quarters before the pandemic) and in CPI OER (full sample) with new-tenant rents' change some quarters earlier
How far ahead new-tenant rents lead
Correlation of CPI y/y in quarter t with R-CPI-NTR y/y in quarter t − k. Full sample from 2006 (83 quarters at the peak lead); sensitivity on quarters to 2019. Peak for rent at 4 quarters (full) and 3 (pre-2020).
Source: the Shelter and rents page; BLS research rent indexes and CPI · Pearson correlation of four-quarter log changesthemes.housing_supply-19

What the rule says CPI rent does next — carried forward quarter by quarter from the last new-tenant edition BLS published, with its band and the quarters already published — is read on the Inflation nowcast page, beside the completions the pipeline implies.

Four tests of the view, none crossed

What would change the view?

Exhibit 18 · What would change the view, test by test

Thresholds from the September report; status from the latest data. 1 crossed · 0 closing · 1 not crossed · 0 pending · 2 not tracked here.
About
StatusTestLatest readingThresholdIf crossedThreshold source
crossedCPI shelter above 3.5% at a three-month annualised rate3.90% (Oct 2026); on 11 Sep 3.22% (Sep 2026)Shelter and rents> 3.50%Shelter is re-accelerating now, not in 2028; the buyer's-market reading (demand at 6.8% mortgages, shelter falling through 2027) fails.the September 2026 report · §1.6c falsifier; §4.4 #13 · 13 Sept 2026
not tracked hereApartment List national rents above +2% y/y by Decemberby December 2026Apartment List (subscription) — quoted, not tracked; −0.8% y/y in August 2026, vacancy 7.1% (issue)–Market rents are rising before the delivery cliff; the pipeline disinflation into 2027 fails.the September 2026 report · §1.6c falsifier; §4.4 #13 · 13 Sept 2026
not tracked hereExisting-home months' supply (NAR) below 4.0NAR existing-home months’ supply is subscription data and carried by no page of this site — quoted, not tracked; 4.9 in August 2026 (issue)< 4.0 monthsThe buyer's market has closed; the demand-collapse reading of housing fails.the September 2026 report · §1.6c falsifier; §4.4 #13 · 13 Sept 2026
not crossedMonths' supply of new homes below 8.0 by the first quarter of 2027by Q1 20278.4 months (Sep 2026)New-home supply< 8.0 monthsThe housing objection to a sequence is removed; 2022, not 1973, is the right analogue.the September 2026 report (as published on 15 Sep) · §4.5 falsifier · 15 Sept 2026
Source: the September 2026 report (thresholds); the pages named in each row (readings)themes.housing_supply-14

The paths the report named, as stages on a calendar

Where does shelter go in 2027–28, and which past episode does this resemble?

The report's named paths for shelter inflation in 2027–28

The September report's dated outlook for CPI shelter and the path that would falsify it, each named, with no odds
About
  1. Down through H1 2027 — the pipelineLagged flat 2025 market rents keep the twelve-month shelter rate drifting below 3% into 2027: a one-to-one-and-a-half-year lag from market rents (Dallas Fed) and about 40% annual pass-through (Cleveland Fed), whatever the Committee does.
  2. Re-acceleration as the delivery trough arrives — H2 2027 into 20282027 deliveries are the cycle low; the supply that holds rents down stops arriving in 2028, and a hike's tenure effect lands in the same year.
  3. Re-acceleration now — the falsifier pathCPI shelter above 3.5% at a three-month annualised rate, or Apartment List rents above +2% y/y by December: the pipeline disinflation is not happening.

The September report states the balance of risks by horizon — down through H1 2027, re-acceleration risk from H2 2027 — without a probability for either; none is shown.

About this exhibit
A text card: the named paths for CPI shelter (down through the first half of 2027; re-acceleration from the second half into 2028; re-acceleration now, the falsifier), with no probabilities
Where shelter inflation goes, 2027–28
The September report’s dated outlook for CPI shelter and the path that would falsify it
Source: research note: Inflation sources outlook (13 Sep 2026) · §3 trend and outlook ('Balance') · 13 Sept 2026themes.housing_supply-scenario-shelter_path

Kinds of past tightening episode, and the test between them

The September report's test across past supply-shock episodes: was new-home months' supply rising or falling into the rate decision?
About
  1. 1973 and 2006 — housing already turning; the tightening was reversedNew-home months' supply high and rising into the decision (9.4 in October 1973, 6.3 in June 2006); starts fell 40.5% after 1973 and 19.6% after 2006.
  2. 2022 — housing absorbed the tighteningMonths' supply fell from 9.4 to 7.5 over the following year with national prices flat. The September report’s test: new-home months' supply below 8.0 by Q1 2027.

The September report argues the present configuration resembles the reversed episodes on this one dimension; it attaches no probability to the analogue. None is shown.

About this exhibit
A text card: past tightening episodes in two groups, told apart by new-home months' supply (housing already turning, tightening reversed; housing absorbed it), and the test between them
Which tightening episode housing resembles at the decision
The September report’s discriminator across eight supply-shock episodes
Source: the September 2026 report (as published on 15 Sep) · §4.5 What the record says; falsifier · 15 Sept 2026themes.housing_supply-scenario-analogue

One rate hike or several: when each one's rent effect lands

What each named policy path does to shelter through 2027–28, as the September report argued · no odds given
About
  1. One move, framed as a recalibrationThe shelter effect of a single hike peaks in twelve to eighteen months through the tenure channel (a 25bp shock raises rents about 0.4–0.5% at six to twelve months, Dias–Duarte) and lands in the same year as the delivery cliff.
  2. A sequence'A sequence lands four of them there': the tenure channel inverted four times into a 2028 delivery cliff, and far harder to abandon.

The logged FOMC-row probabilities for 16 September (recalibration 55 / cycle 31 / hold 14) belong to the decision and are shown on The decision vs the source; the September report attaches no probability to their housing consequences, so none is shown here.

About this exhibit
A text card: the policy paths (one hike framed as a recalibration, or a sequence of hikes) and when each one's rent effect lands against the delivery low, with no probabilities
One move or a sequence into the 2028 cliff
What the policy path does to shelter by 2028
Source: the September 2026 report (as published on 15 Sep) · §4.5 The longer view; What §4.5 concludes · 15 Sept 2026themes.housing_supply-scenario-path

What would each path move?

: the two policy paths this thesis weighs, as the Scenarios set prices them, and what each would do to mortgage rates

Exhibit 19 · Each policy path's 10-year yield, mortgage rate and S&P 500

Model-conditional: each path's 10-year, against 5.12% on 16 Nov 2026; mortgage-rate change at a year, with band; S&P 500, premium and growth held
One move, framed as a recalibrationthe Hold path on the Scenarios set · one move, then the hold path
The 10-year this path justifies4.07%−106bp from 5.12% on 16 Nov 2026
Mortgage rate a year on−101bp90% band −120 to −83bp

S&P 500 +29.6% (9,610) if the premium and growth were unchanged — the fair-value model's arithmetic on that 10-year, not a forecast.

A sequencethe SEP path on the Scenarios set · the SEP's further move: a 2026 median of 4.1% (16 Sep 2026)
The 10-year this path justifies4.12%−101bp from 5.12% on 16 Nov 2026
Mortgage rate a year on−97bp90% band −114 to −79bp

S&P 500 +27.8% (9,477) if the premium and growth were unchanged — the fair-value model's arithmetic on that 10-year, not a forecast.

The same policy paths as on The decision vs the source of inflation and The hike's transmission, under this page's names for them.

The channel these paths run through · The 10-year → the new-purchase mortgage rate: +100bp on the 10-year (monthly average), sustained → +76bp at 3 months, +96bp at 12, 76% passed through in a quarter, 96% in a year; level beta 1.01; 90% band 79 to 113bp. Record: 1998–2011 110bp; 2012–2025 92bp; survey of 1973–2019 68bp (at 12 months). The transmission table carries it with its sample and its model.

What each path would do to core PCE, unemployment and real GDP is read on Thematic impact, under Outlook.

About this exhibit
For each policy path (one hike then a hold; the SEP path): the 10-year yield the path justifies against the anchor, the mortgage-rate change a year on with its band, and the S&P 500 level at an unchanged premium
What would each path move?
Model-conditional: the ten-year column is the fair value the rules grid gives each policy path at today's term premium (What is priced); the S&P 500 column is the valuation identity's answer to that ten-year with the premium and near-term growth held (Scenarios); the mortgage column is the ten-year's move through the transmission table's pass-through at a year, with its band (Transmission). They are sensitivities with their models named, not positions or targets.
Source: the Scenarios and Transmission pages; Macro Review calculations · data to 16 Nov 2026 and Oct 2026themes.housing_supply-15

The research indexes

What BLS published, quarter by quarter, and what it revised

Exhibit 20 · New-tenant and all-tenant research rent indexes, by quarter

Index levels and % changes, quarterly to Q3 2025, with BLS's published change and its interval · below: revisions since the previous edition
QuarterNew tenant rent (R-CPI-NTR)All tenant regressed rent (R-CPI-ATR)
Indexq/q ann.2q ann.y/yPublished (95% interval)Indexq/q ann.2q ann.y/yPublished (95% interval)
Q3 2025215.6+13.9+3.2+1.3+1.2 (-4.0 to 6.5)208.7+2.3+3.0+3.0+2.9 (2.7 to 3.2)
Q2 2025208.7−6.6−0.8−1.9−1.9 (-4.6 to 0.7)207.6+3.7+3.3+3.3+3.3 (2.9 to 3.6)
Q1 2025212.3+5.3−0.6+1.6+1.5 (-1.3 to 4.4)205.7+2.9+3.0+3.5+3.4 (3.0 to 3.9)
Q4 2024209.5−6.2−3.0+1.2+1.2 (-0.4 to 2.8)204.2+3.1+3.4+3.4+3.3 (3.0 to 3.7)
Q3 2024212.9+0.2+3.8+1.7+1.7 (0.4 to 2.9)202.7+3.7+4.0+3.9+3.9 (3.5 to 4.2)
Q2 2024212.8+7.4+5.7+2.4+2.4 (0.8 to 4.0)200.9+4.3+3.4+4.3+4.2 (3.7 to 4.7)
Q1 2024209.0+4.0−0.4+2.7+2.6 (0.4 to 4.9)198.7+2.5+3.9+4.8+4.7 (4.3 to 5.0)
Q4 2023207.0−4.5−0.8+2.4+2.4 (0.6 to 4.2)197.5+5.3+5.1+5.5+5.3 (5.0 to 5.7)

Revisions between editions

Quarter2025q2 edition2025q3 editionRevision, pp
Q1 2024+2.7%+2.6%−0.1
Q2 2024+2.3%+2.4%+0.1
Q3 2024+1.8%+1.7%−0.1
Q4 2024+1.0%+1.2%+0.2
Q1 2025+1.0%+1.5%+0.5
Q2 2025−9.3%−1.9%+7.4
About this exhibit
A table of BLS's new-tenant and all-tenant research rent indexes for the last quarters published: index level and percent changes (quarter and two quarters annualised, year on year, published with its interval), and below it the revisions between editions
Research rent indexes: the quarterly ladder
Index 2000q1 = 100. Changes in % (q/q and two-quarter annualised, y/y as a percent change); the published four-quarter change is BLS's log difference with its 95% interval. Last eight quarters.
Source: U.S. Bureau of Labor Statistics research indexes R-CPI-NTR and R-CPI-ATR; rent rule calculated on this site. From Shelter and rents (data to Jul 2026)themes.housing_supply-21

Timeline and the report's exhibits

Dated events, and the report's exhibits mapped to live pages

Exhibit 21 · The dated events behind the view

Events with their source; scheduled dates are marked, and events read from a data series say so
  • Oct 2025
    Lapse in appropriations: the October 2025 CPI Housing Survey is not collected.

    research note: Hike or hold markets (13 Sep 2026) · §3.6 (BLS verbatim) · 13 Sept 2026

  • 24 Apr 2026
    BLS pauses publication of the New Tenant Rent and All Tenant Regressed Rent indexes; the latest edition is 2025 Q3.

    research note: Hike or hold markets (13 Sep 2026) · §3.6 · 13 Sept 2026

  • 18 Aug 2026
    Census July construction: 5+ unit completions 329,000 a year, −25.6% y/y. Zillow July rents +2.3% y/y, 'the fastest annual pace in over a year'.

    research note: Hike or hold markets (13 Sep 2026) · §3.6 · 13 Sept 2026

  • 25 Aug 2026
    Yardi Matrix Q3 supply forecast: 2027 multifamily deliveries 444,343 units, the cycle low.

    research note: Hike or hold markets (13 Sep 2026) · §3.6 · 13 Sept 2026

  • 8 Sept 2026
    Zillow August rents +2.5% y/y, 'reaccelerating'.

    research note: Inflation sources outlook (13 Sep 2026) · §3 · 13 Sept 2026

  • 10 Sept 2026
    Redfin: 57.9% more sellers than buyers in August — 'the strongest national buyer's market in our records'.

    the September 2026 report · §1.6c key points · 13 Sept 2026

  • 11 Sept 2026
    August CPI: shelter 3.0% y/y and 2.1% annualised over three months, after 4.9% over March–May.

    research note: Inflation sources outlook (13 Sep 2026) · §3; headline finding 4 · 13 Sept 2026

  • 13 Sept 2026
    Issue published: the shelter perversity is real on a two-year view, not a two-quarter view.

    the September 2026 report · §1.6c analysis · 13 Sept 2026

  • 15 Sept 2026
    the report as published on 15 September: 'a sequence lands four of them there'; new-home months' supply below 8.0 by Q1 2027 set as the housing falsifier.

    the September 2026 report (as published on 15 Sep) · §4.5 The longer view; falsifier · 15 Sept 2026

  • 17 Sept 2026
    On the September report’s watch list: August housing starts and Lennar's third quarter — 'the 2028 supply problem'.

    the September 2026 report · §4.3 What to watch · 13 Sept 2026

  • 17 Sept 2026
    Interest on reserve balances moves to 3.90% (+25bp) from 17 Sep 2026: the 16 September decision implemented — the tenure channel's clock starts. data

    Policy rates · series iorb (Board PRATES) · 17 Nov 2026

  • Dec 2026
    Window for the Apartment List falsifier (+2% y/y by December) closes. scheduled

    the September 2026 report · §1.6c falsifier · 13 Sept 2026

  • Q1 2027
    Window for the months' supply falsifier (new homes below 8.0) closes. scheduled

    the September 2026 report (as published on 15 Sep) · §4.5 falsifier · 15 Sept 2026

Source: the September 2026 report and the pages cited per rowthemes.housing_supply-16

Exhibit 22 · The September report's housing exhibits and the pages that now carry them

Each named exhibit as the September report printed it, fixed at its 11 Sep 2026 data date, with the page that now carries it live
No.StemExhibitLive page
49E14c_rents_vs_cpi_shelterPrivate rent indices vs CPI shelter, latest twelve-month changeShelter and rents
51E14h_pipelineStarts, completions and units under construction, 2015–2026Construction pipeline
50E14d_mf_delivery_cliffMultifamily deliveries, 2024–2027quoted figures — no live page
153E24c_months_supply_at_decisionMonths' supply of new homes at the moment of each decisionNew-home supply
48E14b_months_supplyMonths' supply of existing homes, 2012–2026quoted figures — no live page
52E14e_builder_incentivesBuilder sentiment, price cuts, incentives and new-home supplyquoted figures — no live page
47E14_sellers_vs_buyersRedfin: sellers and buyers in the market, August 2026quoted figures — no live page
55E14i_mortgage_rateThirty-year fixed mortgage rate, 2019–2026quoted figures — no live page
53E14f_metro_pricesHouse-price change: national indices and the metros in declinequoted figures — no live page
22E2r_source_outlookSource by source: driver, dated outlook and channelSources & rate sensitivity
155E24e_2027_outlookThe published forecasts for 2027, and what they imply against todayquoted figures — no live page
About this exhibit
A table: each housing exhibit in the September report (number, title) and the live page that now carries it, or 'quoted figures — no live page'
The September report's exhibits, and where each lives now
Exhibit numbers as the report printed them on 15 September; each stays fixed at the 11 September data date and points to the page that carries it live
Source: the September 2026 report, exhibit listthemes.housing_supply-17

Method

References, the housing pages, the rent rule, the status rule and the terms65 checks, all pass
About this page

The September report argued that a hike does not raise shelter CPI in 2027 — the construction pipeline does the opposite — but raises the 2028 problem, when the multifamily delivery low arrives and the shift from owning to renting a hike causes lands on it. This page keeps that argument against the shelter, national-accounts and housing pages and scores the report's tests.

Exhibit 23 · Definitions

  • Theme
    What the page reads.

    The theme follows the shelter pipeline from the building that is not happening to the rents that will be charged: CPI shelter and its components and the paused new-tenant index from the CPI's shelter series, residential investment from the GDP & GDI page, and the Census construction pipeline and months' supply from the Construction pipeline and New-home supply pages. Every live value is a copy; nothing is recomputed.

  • Housing
    Pages computed elsewhere.

    This page reads the Construction pipeline and New-home supply pages on every update and finds their series by label (permits, starts, completions, under construction; total and 5+ units; months' supply, new and existing). Until they are published their trackers and triggers are pending, and the September report’s figures appear only as quotations. When they are published, the September report’s July 2026 figures are checked against them; if a later release has revised July, the revision is shown rather than treated as a failure.

  • Rule
    New tenant rents to CPI rent.

    The rule fits CPI rent's four-quarter change on the New Tenant Rent Index four quarters earlier (full sample) and projects one year ahead from the last edition. It is a correlation of changes, not a structural model, and the input series is paused.

  • Lead
    How the lead is measured and the rule scored.

    The lead is measured, not assumed: CPI rent's four-quarter log change in quarter t is correlated with the new-tenant index's four-quarter change in quarter t − k for k of nought to eight quarters, on the full sample from 2006 and again on the quarters to 2019, and the peak of that curve is the lead the rule uses. The rule is then ordinary least squares of the one on the other at that lead, fitted before 2020 and scored after it against carrying rent's own change forward — the comparison that says whether the index adds anything. Projections use the full-sample fit with a ±1.96σ residual band, which ignores parameter uncertainty and the revisions the newest new-tenant quarters still carry.

  • Revisions
    Why the newest quarters move.

    A repeat-rent index learns about a quarter only when the next tenant moves in, so BLS re-estimates the whole series each edition and its newest quarters rest on the fewest pairs. The quarters that moved between the 2025q2 and 2025q3 editions are printed as they moved; the site does not revise them.

  • Status
    How triggers are scored.

    crossed — the threshold condition holds on the latest observation (for a run, on each of the last n); closing — not crossed, the series has printed since the September report's data date (11 September 2026), and the gap to the threshold is half or less of the gap on that date; not crossed — otherwise; pending — the observable is dated in the future or no page of this site carries it yet; not tracked here — an event, or subscription data this site does not display. Colours flag attention, not good or bad news for the book.

  • Odds
    None given.

    The September report gives no probabilities for this theme. The scenario sets name the dated outlook, the analogue and the policy path without odds; the decision's logged FOMC probabilities are on the decision page.

  • Projection
    Completions implied by starts already made.

    The multifamily and single-family completions drawn after the latest month are the Construction pipeline page's distributed-lag model, copied here with its band: completions on current and past starts with Almon weights fitted 1990–2019, starts after the latest month held at their three-month average, and a 90% band from the model's own out-of-sample misses. It is model-conditional — what starts already made imply if builders keep their past pace — not a forecast of deliveries.

  • Verdict
    How the verdict word is set.

    The verdict word is set by rule on every update: resolved — the log records the thesis as resolved; tested — a trigger of the falsifier table has crossed and the log has not yet closed the theme; under pressure — the log marks the theme closing, or a falsifier trigger is closing; holding — otherwise. The words of the verdict restate a passage of the logged claim, which this update checks is in the log.

  • Terms
    The words of the view.

    · · · .

  • Documents
    Where the view was argued.
    • the September 2026 report · §1.6c Housing — the chain, link by link · 13 Sept 2026
    • the September 2026 report · §4.5 The longer view — shelter runs the wrong way · 13 Sept 2026
    • research note: Hike or hold markets (13 Sep 2026) · §0.7 and §3.6 The shelter pipeline into 2027 · 13 Sept 2026
    • research note: Inflation sources outlook (13 Sep 2026) · §3 Shelter and housing services — trend and outlook · 13 Sept 2026
    • research note: Hike or hold case (13 Sep 2026) · §0.6 and §4.3 Months' supply at the decision · 13 Sept 2026
    • the September 2026 report (as published on 15 Sep) · §4.5 The longer view; falsifier · 15 Sept 2026
  • Licences
    What is quoted only.

    CPI, the research rent indexes, the national accounts and Census construction data are published; the rent rule is calculated on this site. Apartment List, Zillow, RealPage, CoStar, Yardi Matrix, CBRE, NAR, Redfin and NAHB are proprietary and appear only as quotations from the September report, with dates.

Source: Macro Reviewthemes.housing_supply-18

Before you read

Macro Review is a personal research site, still being built.

Nothing on it is investment advice, a recommendation or a solicitation.

Check every figure for yourself. Each names its publisher — confirm it there before you rely on it.

Figures can be revised, delayed or wrong.

Cones, fans, scenarios, fair values and “what followed” tables are conditional on the model each page names: they describe, they do not predict.

No warranty is given; use is at your own risk.