Skip to content
STACKX RESEARCHHOUSING CYCLE / 01

Charlotte · Raleigh · Dallas–Fort Worth · Houston · Phoenix · Tampa

Fewer permits. A recovery—or a warning?

Construction authorizations are falling in five of six metros. Yet payrolls and the home-price index still grew in all six. The useful signal is in the mix—and in what these measures cannot resolve.

Our reading of the evidence

Future supply restraint is visible. Rental recovery is not established.

Fewer authorizations can eventually mean fewer competing homes. They can also reflect weaker expected leasing, tighter financing or a high prior-year base. Positive payroll growth and positive annual price changes challenge a simple collapse narrative, but neither establishes current rental pricing power.

The investor implication: separate the later supply outlook from the rent and occupancy achievable today. A purchase that needs a rent rebound remains a recovery bet until local leasing evidence supports it. This report identifies where that distinction is most consequential.

Inspect the comparison ↓
01 / THE MIX

Charlotte’s total falls 17.7%; its 5+ category rises 9.7%.

A headline supply reduction conceals different directions by building size.

02 / THE TENSION

Raleigh pairs 2.3% job growth with 10.5% more permits.

Stronger employment context comes with potential new competition, plus material imputation.

03 / THE OPEN QUESTION

Phoenix cuts authorizations 18.7%; jobs grow 0.9%.

Later supply relief and a response to softer demand remain competing explanations.

Calculated from BLS payroll series via FRED ↗ and Census July 2026 CBSA YTD permits ↗, compared with July 2025 YTD ↗. Exact metro sources and formulas follow.

01 / Common-period comparison

The same question, six different answers.

The six metros are a defined research sample, not a national ranking. Employment describes July; prices describe the second quarter; permits cover the first seven months. Compare each indicator with its own prior-year period, rather than treating them as simultaneous observations.

Six-metro evidence sheet · percentages are changes, not investment scores
MetroPayroll jobs
Jul ’26 YoY · SA
Home prices
Q2 ’26 YoY · NSA
All permitsSingle-unit5+ units
Charlotte ↓+1.2%+1.5%-17.7%-27.6%+9.7%
Raleigh ↓+2.3%+1.1%+10.5%+2.8%+25.1%
Dallas–Fort Worth ↓+1.1%+0.6%-6.3%-3.3%-8.3%
Houston ↓+1.2%+3.4%-12.5%-5.0%-36.6%
Phoenix ↓+0.9%+0.8%-18.7%-12.0%-33.1%
Tampa ↓+0.3%+0.5%-3.9%+0.3%-8.6%

Permit columns: housing units authorized, January–July 2026 versus January–July 2025, not seasonally adjusted. Single-unit and 5+ omit 2–4-unit buildings; all permits includes them. Full metropolitan areas throughout.

Payroll: BLS via FRED, August 21 release. Prices: FHFA August 25 release ↗. Permits: Census BPS, August 25 release. Sources are linked for each metro below.

02 / Historical perspective

A longer lens on today’s change.

The employment and price charts start in 2019. Each metro is rebased to the same starting value, so the lines compare change over time. They do not compare dollar prices, affordability, rental returns or current investment value.

Explore the evidence

Compare the paths, not just the latest number.

Metros shown

Employment histories are rebased to January 2019 = 100 to compare change across metros of different sizes. The source series measures seasonally adjusted payroll jobs in thousands; the chart shows relative change, not the number of households or housing demand.

Showing Charlotte, Raleigh, Dallas–Fort Worth, Houston, Phoenix, Tampa.

Latest observations in the selected series

CharlotteJul 2026: 1,402.8 thousand jobs113.0 on the rebased chart · base Jan 2019

RaleighJul 2026: 784.1 thousand jobs122.1 on the rebased chart · base Jan 2019

Dallas–Fort WorthJul 2026: 4,362.6 thousand jobs117.4 on the rebased chart · base Jan 2019

HoustonJul 2026: 3,508.3 thousand jobs112.0 on the rebased chart · base Jan 2019

PhoenixJul 2026: 2,484.8 thousand jobs115.8 on the rebased chart · base Jan 2019

TampaJul 2026: 1,555.0 thousand jobs113.3 on the rebased chart · base Jan 2019

Inspect the data table Original observations and rebased values
Employment: seasonally adjusted thousands of jobs; chart baseline January 2019 = 100. Missing observations are not filled.
PeriodCharlotteThousand jobs / rebasedRaleighThousand jobs / rebasedDallas–Fort WorthThousand jobs / rebasedHoustonThousand jobs / rebasedPhoenixThousand jobs / rebasedTampaThousand jobs / rebased
Jan 20191,241.4 / 100.0642.1 / 100.03,715.1 / 100.03,133.4 / 100.02,144.9 / 100.01,372.6 / 100.0
Feb 20191,243.7 / 100.2641.1 / 99.83,723.9 / 100.23,142.2 / 100.32,148.2 / 100.21,376.3 / 100.3
Mar 20191,245.6 / 100.3642.3 / 100.03,730.2 / 100.43,140.2 / 100.22,156.5 / 100.51,377.2 / 100.3
Apr 20191,246.7 / 100.4642.5 / 100.13,737.8 / 100.63,145.0 / 100.42,161.7 / 100.81,378.2 / 100.4
May 20191,247.6 / 100.5642.7 / 100.13,749.2 / 100.93,152.4 / 100.62,169.7 / 101.21,381.0 / 100.6
Jun 20191,250.9 / 100.8642.6 / 100.13,755.6 / 101.13,155.3 / 100.72,170.5 / 101.21,381.8 / 100.7
Jul 20191,253.2 / 101.0643.6 / 100.23,774.5 / 101.63,167.9 / 101.12,175.4 / 101.41,385.6 / 100.9
Aug 20191,258.3 / 101.4645.4 / 100.53,784.3 / 101.93,175.5 / 101.32,190.2 / 102.11,388.7 / 101.2
Sep 20191,260.0 / 101.5647.5 / 100.83,790.0 / 102.03,180.6 / 101.52,193.5 / 102.31,392.7 / 101.5
Oct 20191,262.5 / 101.7650.1 / 101.23,802.2 / 102.33,179.2 / 101.52,197.5 / 102.51,394.1 / 101.6
Nov 20191,266.2 / 102.0651.6 / 101.53,822.1 / 102.93,185.7 / 101.72,205.9 / 102.81,398.7 / 101.9
Dec 20191,265.4 / 101.9653.0 / 101.73,824.1 / 102.93,186.3 / 101.72,214.6 / 103.21,401.6 / 102.1
Jan 20201,273.5 / 102.6656.6 / 102.33,830.0 / 103.13,195.4 / 102.02,217.8 / 103.41,405.2 / 102.4
Feb 20201,275.5 / 102.7658.1 / 102.53,835.5 / 103.23,199.7 / 102.12,218.2 / 103.41,409.6 / 102.7
Mar 20201,267.4 / 102.1656.4 / 102.23,814.1 / 102.73,177.5 / 101.42,208.4 / 103.01,398.7 / 101.9
Apr 20201,117.8 / 90.0576.9 / 89.83,397.7 / 91.52,832.2 / 90.41,967.7 / 91.71,232.7 / 89.8
May 20201,140.0 / 91.8587.3 / 91.53,471.3 / 93.42,886.3 / 92.12,035.7 / 94.91,283.3 / 93.5
Jun 20201,177.2 / 94.8606.1 / 94.43,543.4 / 95.42,929.4 / 93.52,086.9 / 97.31,321.7 / 96.3
Jul 20201,195.0 / 96.3614.4 / 95.73,576.7 / 96.32,929.2 / 93.52,069.3 / 96.51,324.0 / 96.5
Aug 20201,204.1 / 97.0623.5 / 97.13,606.2 / 97.12,943.2 / 93.92,095.0 / 97.71,337.2 / 97.4
Sep 20201,214.7 / 97.8630.8 / 98.23,629.8 / 97.72,959.0 / 94.42,109.0 / 98.31,338.2 / 97.5
Oct 20201,235.0 / 99.5637.0 / 99.23,673.3 / 98.92,981.9 / 95.22,129.9 / 99.31,351.7 / 98.5
Nov 20201,239.3 / 99.8640.8 / 99.83,687.5 / 99.32,994.2 / 95.62,142.8 / 99.91,358.7 / 99.0
Dec 20201,243.7 / 100.2645.2 / 100.53,707.2 / 99.83,001.5 / 95.82,151.9 / 100.31,367.5 / 99.6
Jan 20211,242.7 / 100.1646.7 / 100.73,720.5 / 100.13,005.1 / 95.92,160.2 / 100.71,371.1 / 99.9
Feb 20211,245.3 / 100.3648.7 / 101.03,707.7 / 99.82,991.1 / 95.52,168.3 / 101.11,375.2 / 100.2
Mar 20211,248.0 / 100.5653.5 / 101.83,747.0 / 100.93,018.1 / 96.32,180.2 / 101.61,382.2 / 100.7
Apr 20211,251.2 / 100.8658.5 / 102.63,760.2 / 101.23,028.0 / 96.62,182.2 / 101.71,377.6 / 100.4
May 20211,253.8 / 101.0662.9 / 103.23,770.2 / 101.53,034.6 / 96.82,198.5 / 102.51,389.7 / 101.2
Jun 20211,251.5 / 100.8667.7 / 104.03,788.4 / 102.03,046.4 / 97.22,214.9 / 103.31,398.7 / 101.9
Jul 20211,262.9 / 101.7672.6 / 104.83,838.4 / 103.33,091.2 / 98.72,224.6 / 103.71,407.8 / 102.6
Aug 20211,260.0 / 101.5671.8 / 104.63,846.3 / 103.53,096.1 / 98.82,235.4 / 104.21,413.8 / 103.0
Sep 20211,262.9 / 101.7671.3 / 104.53,862.3 / 104.03,107.9 / 99.22,243.7 / 104.61,422.1 / 103.6
Oct 20211,282.1 / 103.3678.1 / 105.63,912.5 / 105.33,143.8 / 100.32,256.0 / 105.21,435.5 / 104.6
Nov 20211,286.3 / 103.6680.5 / 106.03,933.8 / 105.93,157.2 / 100.82,260.8 / 105.41,443.2 / 105.1
Dec 20211,291.6 / 104.0684.1 / 106.53,956.7 / 106.53,173.5 / 101.32,273.7 / 106.01,450.7 / 105.7
Jan 20221,290.3 / 103.9682.8 / 106.33,956.7 / 106.53,164.6 / 101.02,276.6 / 106.11,451.6 / 105.8
Feb 20221,304.5 / 105.1691.0 / 107.63,986.8 / 107.33,184.4 / 101.62,290.6 / 106.81,457.9 / 106.2
Mar 20221,308.2 / 105.4694.4 / 108.13,998.3 / 107.63,193.7 / 101.92,292.6 / 106.91,460.0 / 106.4
Apr 20221,312.4 / 105.7698.1 / 108.74,028.6 / 108.43,218.5 / 102.72,313.4 / 107.91,472.7 / 107.3
May 20221,314.9 / 105.9698.1 / 108.74,037.3 / 108.73,225.7 / 102.92,326.6 / 108.51,477.0 / 107.6
Jun 20221,315.1 / 105.9703.2 / 109.54,044.3 / 108.93,229.5 / 103.12,329.3 / 108.61,479.5 / 107.8
Jul 20221,322.7 / 106.5710.6 / 110.74,096.6 / 110.33,275.4 / 104.52,341.6 / 109.21,488.7 / 108.5
Aug 20221,324.6 / 106.7711.2 / 110.84,113.8 / 110.73,286.8 / 104.92,347.2 / 109.41,495.9 / 109.0
Sep 20221,327.9 / 107.0711.7 / 110.84,124.4 / 111.03,296.2 / 105.22,351.8 / 109.61,499.1 / 109.2
Oct 20221,328.5 / 107.0711.1 / 110.74,140.0 / 111.43,308.8 / 105.62,355.2 / 109.81,500.5 / 109.3
Nov 20221,329.8 / 107.1713.5 / 111.14,148.1 / 111.73,317.4 / 105.92,362.2 / 110.11,504.2 / 109.6
Dec 20221,331.6 / 107.3714.6 / 111.34,146.4 / 111.63,325.0 / 106.12,361.7 / 110.11,507.9 / 109.9
Jan 20231,339.6 / 107.9717.8 / 111.84,167.6 / 112.23,341.8 / 106.72,368.4 / 110.41,513.8 / 110.3
Feb 20231,342.5 / 108.1720.5 / 112.24,170.5 / 112.33,348.7 / 106.92,374.1 / 110.71,515.8 / 110.4
Mar 20231,344.4 / 108.3722.4 / 112.54,181.7 / 112.63,356.2 / 107.12,374.7 / 110.71,518.5 / 110.6
Apr 20231,344.3 / 108.3721.2 / 112.34,186.4 / 112.73,360.2 / 107.22,394.6 / 111.61,523.6 / 111.0
May 20231,347.1 / 108.5724.0 / 112.84,190.1 / 112.83,364.8 / 107.42,402.6 / 112.01,524.3 / 111.1
Jun 20231,353.8 / 109.1732.2 / 114.04,200.1 / 113.13,376.0 / 107.72,413.2 / 112.51,529.7 / 111.4
Jul 20231,353.5 / 109.0730.6 / 113.84,208.6 / 113.33,379.7 / 107.92,414.9 / 112.61,532.6 / 111.7
Aug 20231,354.9 / 109.1732.3 / 114.04,215.0 / 113.53,388.8 / 108.22,423.3 / 113.01,535.4 / 111.9
Sep 20231,356.3 / 109.3733.3 / 114.24,218.5 / 113.63,401.3 / 108.52,431.8 / 113.41,535.4 / 111.9
Oct 20231,357.4 / 109.3735.6 / 114.64,216.6 / 113.53,399.7 / 108.52,400.7 / 111.91,538.2 / 112.1
Nov 20231,358.9 / 109.5737.2 / 114.84,224.6 / 113.73,399.8 / 108.52,413.4 / 112.51,540.4 / 112.2
Dec 20231,361.5 / 109.7742.8 / 115.74,227.5 / 113.83,406.4 / 108.72,424.6 / 113.01,540.7 / 112.2
Jan 20241,359.4 / 109.5741.3 / 115.44,236.7 / 114.03,410.0 / 108.82,448.5 / 114.21,543.8 / 112.5
Feb 20241,362.4 / 109.7742.8 / 115.74,239.6 / 114.13,414.6 / 109.02,452.2 / 114.31,544.5 / 112.5
Mar 20241,363.2 / 109.8745.3 / 116.14,240.8 / 114.23,422.4 / 109.22,455.6 / 114.51,545.5 / 112.6
Apr 20241,366.0 / 110.0747.3 / 116.44,244.2 / 114.23,425.0 / 109.32,453.6 / 114.41,545.9 / 112.6
May 20241,367.0 / 110.1749.3 / 116.74,251.8 / 114.43,429.0 / 109.42,454.1 / 114.41,547.3 / 112.7
Jun 20241,367.7 / 110.2749.0 / 116.64,256.6 / 114.63,433.8 / 109.62,453.1 / 114.41,549.0 / 112.9
Jul 20241,366.9 / 110.1749.3 / 116.74,263.0 / 114.73,409.2 / 108.82,453.8 / 114.41,547.6 / 112.7
Aug 20241,368.3 / 110.2751.1 / 117.04,271.4 / 115.03,448.9 / 110.12,456.0 / 114.51,547.3 / 112.7
Sep 20241,370.9 / 110.4751.9 / 117.14,276.5 / 115.13,455.3 / 110.32,460.7 / 114.71,550.9 / 113.0
Oct 20241,373.1 / 110.6752.5 / 117.24,283.1 / 115.33,459.7 / 110.42,459.7 / 114.71,517.0 / 110.5
Nov 20241,374.0 / 110.7755.5 / 117.74,289.5 / 115.53,463.1 / 110.52,465.4 / 114.91,548.5 / 112.8
Dec 20241,374.7 / 110.7753.8 / 117.44,296.1 / 115.63,465.9 / 110.62,464.7 / 114.91,548.5 / 112.8
Jan 20251,377.0 / 110.9758.2 / 118.14,290.0 / 115.53,466.8 / 110.62,467.0 / 115.01,553.4 / 113.2
Feb 20251,376.8 / 110.9758.2 / 118.14,296.2 / 115.63,468.3 / 110.72,465.2 / 114.91,555.4 / 113.3
Mar 20251,378.1 / 111.0759.1 / 118.24,301.0 / 115.83,468.2 / 110.72,466.3 / 115.01,557.0 / 113.4
Apr 20251,380.1 / 111.2762.7 / 118.84,309.9 / 116.03,476.7 / 111.02,464.8 / 114.91,555.7 / 113.3
May 20251,382.4 / 111.4765.1 / 119.24,315.1 / 116.23,473.2 / 110.82,462.3 / 114.81,555.4 / 113.3
Jun 20251,383.9 / 111.5764.6 / 119.14,304.6 / 115.93,468.7 / 110.72,461.3 / 114.81,551.0 / 113.0
Jul 20251,385.7 / 111.6766.8 / 119.44,316.2 / 116.23,467.5 / 110.72,462.2 / 114.81,549.7 / 112.9
Aug 20251,386.4 / 111.7767.1 / 119.54,316.4 / 116.23,469.7 / 110.72,459.3 / 114.71,548.9 / 112.8
Sep 20251,385.3 / 111.6769.6 / 119.94,318.6 / 116.23,469.0 / 110.72,456.2 / 114.51,547.7 / 112.8
Oct 20251,386.4 / 111.7767.4 / 119.54,308.0 / 116.03,466.8 / 110.62,456.6 / 114.51,545.4 / 112.6
Nov 20251,389.5 / 111.9769.2 / 119.84,311.9 / 116.13,473.9 / 110.92,459.0 / 114.61,547.3 / 112.7
Dec 20251,391.7 / 112.1770.2 / 120.04,319.6 / 116.33,479.9 / 111.12,461.2 / 114.71,549.2 / 112.9
Jan 20261,390.7 / 112.0771.8 / 120.24,323.5 / 116.43,483.0 / 111.22,466.7 / 115.01,549.0 / 112.9
Feb 20261,389.2 / 111.9771.8 / 120.24,319.9 / 116.33,483.3 / 111.22,478.8 / 115.61,550.7 / 113.0
Mar 20261,391.5 / 112.1773.4 / 120.44,340.8 / 116.83,494.4 / 111.52,477.8 / 115.51,552.0 / 113.1
Apr 20261,397.4 / 112.6780.6 / 121.64,343.9 / 116.93,492.5 / 111.52,482.2 / 115.71,556.2 / 113.4
May 20261,399.8 / 112.8783.9 / 122.14,348.6 / 117.13,498.4 / 111.62,481.9 / 115.71,555.4 / 113.3
Jun 20261,400.4 / 112.8783.0 / 121.94,361.8 / 117.43,507.7 / 111.92,483.5 / 115.81,554.0 / 113.2
Jul 20261,402.8 / 113.0784.1 / 122.14,362.6 / 117.43,508.3 / 112.02,484.8 / 115.81,555.0 / 113.3

A small annual change can follow a large cumulative rise.

By Q2 2026, the all-transactions index stood 47.5% above Q4 2019 in Houston and 77.0% above it in Charlotte. Those are nominal changes; they exclude inflation adjustment, operating income, financing and transaction costs. The chart uses Q1 2019 as its common starting quarter.

This price measure has a specific lens.

FHFA’s all-transactions series includes refinance appraisals alongside sales. Its coverage centers on single-family mortgage transactions acquired or securitized by Fannie Mae and Freddie Mac. It is useful historical context, but a purchase-only index or current property comps can tell a different short-term story.

03 / Local interpretation

What deserves investigation in each metro.

Each thesis separates observation, a possible mechanism, contrary evidence and the next property-level check. Watch conditions are editorial review rules, not validated trading signals or forecasts. For payroll rules, two releases means two consecutive observation months evaluated in the latest available vintage, not two revisions to the same month. Successive year-to-date permit releases overlap; they are not independent confirmations. These conditions are not monitored automatically.

01

Charlotte-Concord-Gastonia, NC-SC · CBSA 16740

Total permits fall, but five-plus-unit authorizations rise.

+1.2%Jul payroll YoY+1.5%Q2 HPI YoY12,211Jan–Jul units authorized+9.7%5+ units vs Jan–Jul ’25

Charlotte’s authorization pullback is concentrated in single-unit structures. The five-plus-unit category is expanding even as the total falls. A rental acquisition thesis based only on the total permit decline would miss that split.

The mechanism to test

If authorized five-plus-unit projects proceed, they can add competition for renters in their catchment areas after construction. Fewer single-unit authorizations could reduce a different part of future competition. The two effects cannot be netted into a single rent forecast without locations, tenure and delivery dates.

What challenges this reading

A permit is an option to build, not a completed apartment. Projects can be delayed or abandoned, and some five-plus-unit buildings will be condominiums. The payroll expansion supports demand context but does not tell us which households can afford a particular rent.

What to verify before underwriting

For an existing rental, map five-plus-unit projects within its realistic competitive area and record construction stage, expected availability, unit mix and advertised concessions. A metro-wide single-unit slowdown is insufficient support for higher renewal rents.

What would change our view

Revisit the supply split if the five-plus-unit January-to-date change turns negative in two successive releases. A revision that erases the present increase would also weaken this interpretation.

02

Raleigh-Cary, NC · CBSA 39580

Payroll growth and permit expansion are occurring together.

+2.3%Jul payroll YoY+1.1%Q2 HPI YoY12,123Jan–Jul units authorized+25.1%5+ units vs Jan–Jul ’25

Raleigh has the fastest July payroll growth in this six-metro sample and is the only metro with rising total January–July authorizations. That combination supports investigating demand, but it also creates a clear test: can occupied housing grow fast enough where the new units are located?

The mechanism to test

More payroll jobs may support additional housing demand, while the increase in authorized five-plus-unit housing may enlarge the future competitive set. Job counts cannot be converted directly into new renter households; existing residents, commuters and multiple jobholders can account for part of the gain.

What challenges this reading

More than one-third of the current total permit estimate is imputed. This is an official estimate, but the reporting gap makes the precise pace and mix less secure than the headline suggests. Supply can also be concentrated far from a target property.

What to verify before underwriting

Request project-level status from the relevant planning offices and compare lease-up progress with the effective rents used in an acquisition model. Investigate employment locations and pay levels before assuming broad job growth reaches the target renter segment.

What would change our view

Recheck the growth case if payroll growth falls below 1% year over year for two releases while five-plus-unit authorizations remain above the prior year. Separately, revisit the supply thesis if later reporting reverses the permit increase.

03

Dallas-Fort Worth-Arlington, TX · CBSA 19100

A modest pullback still leaves a large authorized volume.

+1.1%Jul payroll YoY+0.6%Q2 HPI YoY39,217Jan–Jul units authorized-8.3%5+ units vs Jan–Jul ’25

Dallas–Fort Worth has the largest January–July authorized volume in this sample despite a decline from 2025. Both single-unit and five-plus-unit authorizations are lower. The important acquisition question is where existing and underway units compete—not whether the broad metro has fewer new permits.

The mechanism to test

A slower rate of authorization can limit later supply additions, but does not remove homes already underway. The scale of the metro makes an absolute permit total a poor measure of tightness without housing stock, household formation and completion data.

What challenges this reading

Positive payroll growth and a still-positive annual home-price change are counterweights to a demand-collapse interpretation. However, an all-transactions index includes refinance appraisals and does not measure new-home incentives, apartment values or a neighborhood’s comparable sales.

What to verify before underwriting

Separate Dallas-side and Fort Worth-side submarkets before selecting comparables. Trace nearby projects from authorization to construction, then use actual lease terms and a current lender quote to test the acquisition.

What would change our view

Escalate demand review if subsequent monthly payroll readings show year-over-year declines in two successive releases. A return to positive five-plus-unit authorization growth would weaken the case for a smaller later competitive pipeline.

04

Houston-Pasadena-The Woodlands, TX · CBSA 26420

The sharper supply change is in larger buildings.

+1.2%Jul payroll YoY+3.4%Q2 HPI YoY35,494Jan–Jul units authorized-36.6%5+ units vs Jan–Jul ’25

Houston’s five-plus-unit authorizations are falling much faster than its single-unit authorizations. This is a more specific signal than a generic construction slowdown: the mix of potential future additions is shifting, while payroll employment is still growing.

The mechanism to test

If fewer authorizations become fewer starts and then fewer deliveries, later competition from larger buildings may moderate. Each link needs verification. Existing lease-up inventory can keep effective rents under pressure during the gap between authorization and delivery.

What challenges this reading

The decline could reflect tighter financing, weaker expected leasing or unusually large projects in the prior period. This dataset does not isolate the cause, and single-unit construction remains substantial. No flood, insurance or physical-risk conclusion follows from these metro indicators.

What to verify before underwriting

Evaluate the target property against today’s available units and concessions first. Use project timing to build an alternative supply scenario, and obtain property-specific insurance, flood and maintenance evidence independently.

What would change our view

Weaken the future-supply-relief thesis if five-plus-unit authorizations return to year-over-year growth; weaken the demand support if payroll growth turns negative in two releases. Both would require revisiting the base case.

05

Phoenix-Mesa-Chandler, AZ · CBSA 38060

Broad supply restraint is visible; absorption is still unproven.

+0.9%Jul payroll YoY+0.8%Q2 HPI YoY19,547Jan–Jul units authorized-33.1%5+ units vs Jan–Jul ’25

Phoenix has the steepest total authorization decline in this six-metro sample, with reductions in both single-unit and five-plus-unit structures. Payroll growth remains positive but below 1%. This combination merits a two-sided interpretation: later supply may ease, or builders may be responding to softer demand.

The mechanism to test

Restrained authorizations can eventually reduce new competition, but weak expected demand can cause the same pattern. The annual home-price index change remains positive, which challenges a simple collapse narrative, but does not establish rental strength. The missing bridge is current occupancy, absorption and effective rent in the relevant submarket.

What challenges this reading

A percentage decline does not show whether the remaining pipeline is small relative to the housing base. Construction already underway can still deliver after authorizations slow. A metro price index cannot validate a bargain on an individual home.

What to verify before underwriting

Build an acquisition case using observed lease terms and a downside with slower occupancy or weaker rent. Ask which dated evidence would justify a recovery assumption before assigning one to the property.

What would change our view

Reconsider a stabilization thesis if payroll employment falls year over year in two releases, or the next quarterly index shows a negative year-over-year price change. Stronger leasing evidence would be needed before treating a permit decline as rental recovery.

06

Tampa-St. Petersburg-Clearwater, FL · CBSA 45300

Limited payroll growth leaves less room for an easy growth thesis.

+0.3%Jul payroll YoY+0.5%Q2 HPI YoY13,038Jan–Jul units authorized-8.6%5+ units vs Jan–Jul ’25

Tampa has the weakest July payroll growth in this sample. Its total authorizations are only modestly below the prior year, and single-unit authorizations are roughly flat. This is a weaker combination for assuming rapid demand-led tightening than a headline about fewer permits might suggest.

The mechanism to test

Slower payroll growth in percentage terms provides a different employment backdrop from faster-growing peers, but retirement migration and non-payroll income can also drive housing demand. These are missing from the employment measure; weak payroll growth alone cannot establish weak household growth.

What challenges this reading

The five-plus-unit authorization decline could moderate some future competition, while migration could offset slow job growth. Neither possibility demonstrates current rental pricing power. Property insurance, assessments and repair exposure require separate evidence.

What to verify before underwriting

Distinguish workforce, retiree and other renter segments. Verify effective rent, turnover, insurance and building-specific costs before assuming annual rent increases will cover expense growth.

What would change our view

The demand concern would weaken if payroll growth rises above 1% year over year in two releases and local leasing evidence improves. A payroll decline while single-unit authorizations stay near or above 2025 would intensify the need to review competition.

04 / From a metro thesis to a property decision

Keep today’s income separate from tomorrow’s possibility.

For an acquisition

Start with documented leases, collections, vacancy and operating costs. Compare effective rents after concessions and fees on a consistent basis. Then build a second case that delays leasing or lowers rent. The permit trend can inform that scenario; it cannot replace the property evidence.

For example, one free month on a 12-month lease at $2,000 advertised rent means $22,000 in first-year scheduled rent, or about $1,833 per month before other adjustments. An advertised-rent comparison that ignores this concession would overstate that first-year figure by about $167 a month.

For development or a later exit

Translate nearby authorized projects into a dated schedule: planned, started, under construction, completed, leasing and stabilized. Verify each stage independently. Overlap with a project’s lease-up matters more than a metro permit total taken alone.

Keep rent growth, operating-cost growth and exit assumptions separate. A lower entry price does not protect returns if insurance, repairs or debt costs rise. This report supplies cycle context; it does not estimate a cap rate, approve a deal or forecast a sale price.

05 / Reproducible research

Data quality is part of the conclusion.

Prepared with AI-assisted analysis of public releases. No proprietary surveys, property inspections or interviews were performed. The saved dataset was retrieved September 5, 2026; historical observations may be revised in later source releases. This edition does not refresh automatically.

Permit imputation changes the confidence in a precise growth figure.

Census estimates include reported units plus imputation for missing reports. The table below derives the nonreported share from those two published totals. It is a reporting-coverage diagnostic, not a margin of error or a probability that the estimate is wrong.

Share of total authorized units supplied through imputation
MetroJan–Jul 2025Jan–Jul 20262026 5+ units only
Charlotte6.3%15.6%2.3%
Raleigh5.0%37.7%57.7%
Dallas–Fort Worth41.4%19.3%1.9%
Houston7.1%6.4%7.7%
Phoenix6.3%5.4%12.0%
Tampa37.4%1.1%0.8%

Raleigh deserves extra checking: 37.7% of its current total and 57.7% of its five-plus-unit estimate are imputed. Dallas–Fort Worth and Tampa also have substantial prior-year reporting gaps. Comparing reported-only totals would not fix the problem because response coverage changed between years.

Employment

Monthly, seasonally adjusted total nonfarm payroll jobs at establishments, in thousands. July 2026 is compared with July 2025 from the same saved FRED CSV retrieval. This measures jobs, not unique workers, resident employment, new households or renter demand. Preliminary data and history can be revised.

BLS methods and revisions ↗

Home prices

Quarterly nominal FHFA all-transactions HPI, not seasonally adjusted, source index 1995 Q1 = 100. Annual change compares Q2 2026 with Q2 2025. Histories use full MSAs. DFW and Tampa come from FHFA’s newly released multi-division metro workbook, rather than substituting a division or splicing Tampa’s discontinued FRED history.

FHFA HPI methodology ↗

Construction authorizations

Census BPS housing units authorized in privately owned buildings, estimates with imputation, not seasonally adjusted. The comparison uses January–July YTD files including each file’s late reports and corrections. It does not sum monthly releases. Annual 2024 and 2025 totals are separately included in the downloads.

Single-unit, 2-unit, 3–4-unit and 5+ categories describe structure size. They do not identify rental tenure, build-to-rent, starts, completions, demolitions or occupied homes.

Census collection and revision methods ↗

Formulas and geography

Annual change = 100 × (current value ÷ prior same-period value − 1). Chart rebasing = 100 × (observation ÷ first 2019 observation). Imputed share = 100 × (estimated units − reported units) ÷ estimated units.

Source CBSA codes and full metro names are retained in the dataset; source titles may vary. These metro indicators are separate from the city-limit ACS profiles elsewhere on StackX. They are not used to calculate a combined score, implied rental yield or jobs-to-housing balance.

Census field layout ↗

Take the evidence with you.

CSV: 918 data rows, including calculated permit totals. JSON: full history, derived metrics, source links and definitions. Free to download without registration.

Edition record & next review

September 5, 2026 · Version 1. Initial thesis and source snapshot. There is no earlier StackX forecast to score. Future editions should retain this version and record which observations changed, whether the stated watch conditions were met, and why a thesis was retained or revised.

Next evidence to inspect: subsequent BLS metro releases ↗, Census permit releases ↗ and FHFA HPI releases ↗, alongside local effective rents, vacancies, completions and household trends. Those missing local measures are necessary to test rental recovery.

Submit a correction ↗ with the metro, period, indicator and original source. Research context, not a recommendation to buy or sell a security or property.