Losing Ground 2026: How Rising Transportation, Housing & Utility Costs Are Squeezing Moderate-Income Households

A True Lifestyle Cost (H+T+U) analysis of the 30 largest U.S. metropolitan areas

A Certihomes report · by Krishna Malyala

Draft for internal review — not for publication. Prepared June 25, 2026. Revision: US-only + rebuilt-U (real per-metro water + garbage, national cell + streaming + broadband) + measured-T.

Lineage. This report recreates and modernizes Losing Ground: The Struggle of Moderate-Income Households to Afford the Rising Costs of Transportation and Housing (Center for Neighborhood Technology and the Center for Housing Policy, October 2012). We retain CNT's core insight — that housing affordability is incomplete without transportation — and extend it in two ways: (1) we add a fourth dimension, Utilities (U), building the H+T+U "True Lifestyle Cost" framework; and (2) where CNT modeled transportation costs from neighborhood form, we now measure them with our proprietary multimodal cost algorithm, which prices real 2026 fares, tolls, parking, and fuel across 24 of the 30 profiled metros. This report is United States only. This is our original analysis, built entirely on our own data assets; it is not a copy of CNT's text or tables.

Executive Summary

Fourteen years after CNT first warned that housing and transportation costs were outrunning incomes, the squeeze on moderate-income households has not eased — it has hardened, and it has spread to a third front: the monthly cost of keeping the lights, the water, and the connection on.

Applying the Housing + Transportation (H+T) Affordability Index to all 927 U.S. metropolitan areas (118.7 million households), layering on a rebuilt 2026 national utility baseline (real per-metro water + garbage, national cell + streaming + broadband, energy by metro/state — built for all 50 states + DC and 47 US metros), and replacing CNT's modeled transportation with our measured multimodal cost engine, we find:

• Headline 1 — The 45% line is now the exception, not the rule. For moderate-income households (near 80% of Area Median Income), the combined housing-and-transportation burden exceeds the standard 45%-of-income affordability threshold in metros that hold 97.4% of all U.S. households. CNT's central 2012 warning has become the near-universal condition.

• Headline 2 — In the 25 largest metros, 24 of 25 are unaffordable for moderate-income households. The average moderate-income H+T burden across the 25 largest metros is 54.3% of income — nine points above the affordability line. Only Washington, DC sits at the 45% edge.

• Headline 3 — Transportation, not housing, is the swing cost. Nationally, moderate-income households spend 23.9% of income on transportation — about three-quarters as much as the 32.5% they spend on housing. In car-dependent metros transportation runs to a quarter of income before a single rent check is written.

• Headline 4 — Adding the "U" pushes the typical burden to ~67% nationally and to 79% in the worst metro. Once the full essential-utility basket is counted — a national $639.38/month, $7,673/year for the typical moderate-income household — the national True Lifestyle Cost (H+T+U) reaches 67.1% of income, and 79.0% in Miami. This revision rebuilds the U basket with real per-metro water and garbage (mapped from per-municipality utility schedules), a national cellular line, and a new streaming line ($35/mo) alongside national broadband. The three line items CNT never counted at all — broadband, cellular, and streaming — together add $237.25/month ($2,847/year); the broadband-plus-cell "connectivity gap" alone is $202.25/month ($2,427/year), a necessity that did not exist in 2012. Across our 47-metro US national build the full basket runs from $539/month in Phoenix to $788/month in Boston, with a US-metro median of $661/month — higher than the prior draft because water is now metered at real per-metro rates and garbage is now a separate line.

• Headline 5 — Where the U bites hardest is where incomes lag and the rebuilt water/garbage/energy bill runs high. The rebuilt basket re-shapes the burden map: cold-climate, high-gas, moderate-wage metros (Cleveland, Cincinnati, Pittsburgh, St. Louis) carry a heavy heating-plus-water bill, and high-water-rate metros (Houston, Seattle, Portland, Washington DC) climb on water alone. The most-burdened metros by True Lifestyle Cost are Miami (79.0%), Riverside (75.8%), Tampa (73.2%), Orlando (73.1%), and Los Angeles (72.8%) — regions where moderate incomes are low and car dependence is near-total. The most expensive coastal metros (NYC 63.6%, SF 56.3%, DC 53.9%) rank as more affordable for their residents because high local incomes and real transit options absorb the cost. This is the same paradox CNT identified in 2012 — and it has intensified.

What's New Since 2012

Methodology

The lineage: CNT's Housing + Transportation Index

CNT's foundational argument is that the traditional "30% of income on housing" rule is misleading because it ignores the transportation cost of where that housing sits. A cheap house far from jobs and transit can be more expensive, all-in, than a pricier home in a location-efficient neighborhood. CNT operationalized this with two thresholds, which we adopt unchanged:

• Housing is affordable at ≤ 30% of household income.
• Housing + Transportation combined is affordable at ≤ 45% of household income.
• We add a third tier for severity: H+T ≥ 60% of income is severely burdened.

CNT defined moderate-income households as those earning 50–100% of Area Median Income (AMI), computed per metro. We use the H+T Index's standard 80%-of-AMI cohort as our moderate-income proxy — the canonical midpoint of CNT's band and the cohort carried natively in the index. Implied moderate-incomes in our data range from ~$54K (Tampa) to ~$101K (San Francisco), consistent with CNT's 2012 metro range scaled forward for inflation.

Vintages — what is 2022, what is 2026 (read this before any number)

This report deliberately combines three layers of different vintage. We label every column:

The H+T ratios are the 2022 affordability structure; the U basket and the measured-T dollars are live 2026. We keep them separate rather than forcing a single vintage so the affordability structure stays comparable to CNT while the cost layers reflect what a household actually pays in 2026.

Our advance #1: the "U" dimension and the True Lifestyle Cost — now national

CNT's housing figure (ACS Gross Rent and Selected Monthly Owner Costs) embeds energy and water/sewer inside the rent/owner-cost line. That was reasonable in 2010. But it understates the modern essential basket in two ways: it buries energy and water where they can't be seen or acted on, and it omits entirely two line items that are now non-negotiable for employment, school, and benefits access — home broadband and cellular service.

We therefore compute an explicit Utilities (U) basket. This revision rebuilds that basket: water and garbage are now real per-metro figures (mapped from per-municipality utility-schedule research onto each metro), a cellular line is carried nationally, and a new streaming line is added alongside national broadband. Energy continues to vary by metro and state. The basket is built nationally for the United States from utility_baselines_national (scope losing_ground_2026, year 2026): 50 states + DC × components, 47 US metros × components, and 1 US-national rollup.

The connectivity gap ($202.25/mo broadband+cell, or $237.25/mo including streaming) is flat across all metros; energy, water, and garbage are what move the per-metro U figure. Water is now the second-largest swing after energy: it ranges from $66/mo (Phoenix) to $210/mo (Washington DC) across the build. Garbage is a separate line that runs $0 where a metro bundles refuse into property taxes or rent (Baltimore, Boston, Cincinnati, Houston, Kansas City, New York, Philadelphia, Pittsburgh, Washington DC in this build) up to ~$60/mo (Los Angeles, Miami). Across the 47-metro US national build, the full U ranges from $538.83 (Phoenix) to $787.83 (Boston), with a US-metro median of $661.22/mo — materially higher than the prior draft because water is now metered at real per-metro rates and garbage is now counted separately rather than folded into a flat water figure. To avoid double-counting against CNT's housing figure (which already embeds energy, water, and refuse in rent/owner cost), we report two views: the full U basket for the True Lifestyle Cost headline, and the incremental connectivity-plus-streaming gap ($237.25/mo) as the cleanest apples-to-apples addition to CNT's 2012 frame.

Our advance #2: measured, not modeled, transportation

This is the central methodological change in this revision, and the real differentiator vs CNT. CNT's transportation costs are regression estimates from seven neighborhood and four household variables. We replace the modeled transport column in our metro tables with measured costs from our proprietary multimodal cost algorithm (sub-second per-origin) that prices actual 2026 trips from every residential H3 cell to the metro CBD: real transit fares (gtfs_fares, incl. transfers), real toll schedules (toll_rates), real CBD parking (cbd_parking_metro), and live fuel (energy_prices).

The measured number is a per-commute, out-of-pocket money cost — one-way median cell→CBD cost × 44 trips/month per commuter. It is not the same quantity as CNT's whole-household transportation total, which is dominated by ~$10–12K/yr of auto ownership (payments, insurance, depreciation). The honest comparison is our measured commute cost against CNT's running cost (gas/VMT + transit); the metro table shows both. Where a measured precompute exists (24 of 30 metros), the measured columns are live; the 6 metros without a precompute (Riverside, San Diego, Cleveland, Cincinnati, Kansas City, Baltimore) retain the H+T Index modeled transport.

Data sources

National Findings

Housing and transportation together exceed the affordability line for nearly every American household's metro

Weighting all 927 metros by household count, the moderate-income household's burden breaks down as (H/T ratios = 2022 ACS):

A median-income household nationally sits right at the edge (45.9%). The moderate-income household — earning ~30% less — is 11 points deeper underwater on H+T alone, exactly the gap CNT documented in 2012 (their figure: 59% vs 48%). Add the rebuilt national utility basket and the typical moderate-income True Lifestyle Cost reaches 67.1% of income.

Share of households living in unaffordable metros

By contrast, only 45.0% of households live in metros where the median-income household crosses the 45% line. The affordability crisis is overwhelmingly a moderate-income crisis.

Income kept pace with neither

The mechanism is unchanged from CNT's 2012 finding: housing and transportation costs have risen faster than moderate incomes. What's new in 2026 is that the third front — utilities, including the broadband/cell/streaming connectivity basket and a now-metered water and garbage bill — adds $539–$788/month depending on metro that simply was not fully on the 2012 ledger.

Metro-Level Findings

Table 1 — The 20 most-squeezed large metros by True Lifestyle Cost (H+T+U)

Moderate-income households (~80% AMI). H and T are 2022-ACS H+T Index ratios. U is the rebuilt 2026 utility basket for that metro (energy + real per-metro water + garbage + national cell/streaming/broadband); U% is U/income. H+T+U = H+T (2022) + U% (2026).

(Full 30-metro table, incl. Philadelphia 63.0%, Chicago 62.6%, Austin 62.0%, Boston 60.8%, Seattle 60.5%, Baltimore 59.6%, Denver 58.5%, Minneapolis 57.7%, San Francisco 56.3%, Washington DC 53.9%, in losing-ground-2026-data.json.)

What the rebuilt U changed. The prior draft carried a flat $68.83 water line and no separate garbage line. With real per-metro water + garbage, the U dollar figure jumps (US-national $517.94 → $639.38/mo) and the ranking now reflects local water rates and refuse fees on top of energy: high-water metros (Washington DC $210 water, Houston/Seattle/Portland $175, San Diego $164, Boston $165, LA $154) and high-garbage metros (LA $59.53, Miami $58.08, San Francisco $48.13) climb; low-water/low-energy Phoenix ($66 water, $539 total) sits at the bottom. Cold-winter, high-gas metros (Cleveland $694, Cincinnati $673, Pittsburgh $651) remain high on energy. Miami still tops the list — now on a high H+T burden and a heavy garbage line.

The affordability paradox. New York (H+T+U 63.6%), San Francisco (56.3%), and Washington (53.9%) — the metros with the highest absolute costs — rank as comparatively affordable for their own moderate-income residents, because high local incomes and genuine transit alternatives (NYC: 24% transit commute share, the only large metro where moderate-income households average one car, not two) hold the percentage down. The deepest squeeze falls on lower-cost, car-mandatory sun-belt metros where a $56K income must still finance two vehicles.

Table 2 — Measured transportation cost (our proprietary algorithm, 2026) vs CNT modeled

One-way median cell→CBD measured money cost and the resulting monthly per-commuter bill. This replaces the old modeled transport column. "Drive (monthly-pass)" amortizes the fixed CBD parking/toll floor for a regular commuter. CNT running = gas/VMT + transit (the comparable commute slice; CNT's total adds ~$10–12K/yr auto ownership).

What the measured engine shows that a model cannot. The drive-vs-transit gap is enormous and metro-specific because it prices the actual toll-and-parking geography: a Philadelphia drive commuter pays $1,287/mo (bridge tolls + $34.98/day parking) against $110/mo on transit; a NYC drive commuter pays $1,162/mo ($35 parking + $9 CRZ) against $330 on the most expensive transit in the country. Where the engine has thin transit coverage (DC at 7.6%, San Antonio/Phoenix/Charlotte/Austin near-zero), the cheapest served option is often still driving — the measured engine surfaces the transit desert directly rather than assuming a network exists. The measured number is the product moat: CNT's regression cannot price the bridge toll, the CRZ, or the specific garage rate at the destination.

Measured ≠ CNT total — read the right comparison. The measured monthly figure is commute out-of-pocket per commuter; it excludes car ownership. Across the 24 measured metros, the median measured transit commute is $106/mo and the median measured drive (monthly-pass) commute is $560/mo. CNT's whole-household transportation total ($12–15K/yr) is dominated by auto ownership and is a different quantity; the apples-to-apples slice is CNT's running cost (gas/VMT + transit), shown above.

How 2026 Cost Drivers Hit These Households

CNT's 2012 model used 17¢/mile fuel and a $5,576/year vehicle. The 2026 reality our engine prices is harsher and more volatile.

Gasoline — regular, per gallon (2026-06-23)

National regular average: $4.63/gal (EIA, latest). A two-car California household paying ~$3,768/yr in fuel is paying roughly 60% more per gallon than a Texas household for the same miles — a direct tax on West Coast moderate-income budgets.

CBD parking — median daily rate, 2026

A moderate-income worker who must drive into a Manhattan or Atlanta core faces $700–$760/month in parking alone — more than the entire national utility basket. This is the cost the measured engine surfaces directly in Table 2's drive column, and it is what makes the "drop a car" lever real.

Tolls & congestion pricing (EZ-Pass, 2026)

The NYC Congestion Relief Zone ($9/day) is a cost category that did not exist in 2012. Stacked on a $16.79 Hudson crossing, a NJ-resident moderate-income driver commuting into Manhattan now pays ~$25.79 each way in tolls before parking — exactly the economics our engine prices into NYC's $44 one-way drive median, and why NYC's measured transit burden stays low relative to driving.

Charts (described for the production site)

The companion losing-ground-2026-data.json drives these:

1. Diverging bar — "The 45% line." Each large metro as a horizontal bar of H (blue) + T (orange) + U (green, rebuilt per-metro), with a vertical reference line at 45%. Nearly every bar overshoots; the U segment is the new overshoot.
2. Scatter — "The affordability paradox." X = absolute H+T cost ($/yr); Y = H+T+U as % of moderate income. NYC/SF/DC are high-X but low-Y; Miami/Riverside/Tampa are low-X but high-Y.
3. Grouped bar — "Measured: drive vs transit." Per metro, measured drive (monthly-pass) cost beside measured transit cost, showing the toll-and-parking gap the model cannot see.
4. Choropleth — US H+T+U burden by CBSA, moderate-income cohort (927 US metros), with the rebuilt per-metro U layer.

Policy & Product Implications

Policy

1. Affordability standards must count transportation and utilities. A "30% of income" housing rule certifies as affordable homes whose all-in True Lifestyle Cost exceeds 60% of income in most large metros. Subsidy and zoning eligibility should be scored on H+T+U, not H alone.
2. The highest-leverage intervention is enabling car-shedding. The NYC data proves a household that drops from two cars to one saves more (~$7,000–$10,000/yr in ownership cost) than any plausible fuel-price or rent intervention. Transit, protected bike infrastructure, and location-efficient housing near jobs are affordability policy.
3. Treat broadband, cellular, and streaming as essential utilities. The $202/month broadband-plus-cell connectivity gap ($237/month including streaming) is regressive and invisible in current housing-cost accounting. Lifeline/ACP-style subsidies belong in the affordability calculus.
4. Sun-belt metros need targeted relief — and high-water/high-energy metros need utility relief. The deepest H+T squeeze is in Miami, Riverside, Tampa, Orlando, and San Antonio; the rebuilt U layer additionally flags high-water metros (Washington DC, Houston, Seattle, San Diego, Boston, where water alone runs $164–$210/month) and cold-winter, high-gas metros (Cleveland $694, Cincinnati $673, Pittsburgh $651) where the utility bill alone is $650–$700/month.

Product

1. H+T+U / True Lifestyle Cost is a differentiated consumer product. Google and TravelTime answer "how long"; we answer "how much, all-in, for this household." The home-finder that ranks listings by True Lifestyle Cost — not list price — is the flagship application.
2. Measured beats modeled — and that is now real, not promised. This revision replaces the modeled transport column with the engine's measured per-origin cost across 24 metros, pricing the actual fares, tolls, CBD parking, and fuel. Sub-second, per-origin measured multimodal cost is something CNT's batch regression cannot do interactively. Closing the last 6 metros' precompute and grading non-NYC transit fares by distance is the remaining build.
3. Two-earner households compound the effect. Our multimodal/multi-user capability lets a household evaluate a location against both commuters' true costs simultaneously.

Data Appendix

Full machine-readable figures: reports/losing-ground-2026-data.json (national aggregates, 30-metro table with H/T/U decomposition, national per-metro U, and measured transport costs).

National summary (moderate-income, ~80% AMI, household-weighted across 927 CBSAs)

Data gaps — status before publication

Closed/improved this revision:

1. ✅ Water and garbage are now PER-METRO. The rebuilt utility_baselines_national carries real per-metro water + garbage (mapped from per-municipality utility schedules), plus a national cellular line and a new streaming line alongside national broadband. Energy continues to vary by metro/state. The old flat $68.83 water assumption and the missing garbage line are closed.
2. ✅ Transport is now MEASURED, not modeled. The metro transport column is our proprietary multimodal cost algorithm's measured 2026 cost (fares/tolls/parking/fuel) for 24 of 30 metros. This is the differentiator vs CNT.
3. ✅ Report scope is now strictly UNITED STATES. The prior draft's Canadian metros and the "Canadian U is a US fallback (tier D)" caveat are removed entirely; all metros, the U build, and the national aggregates are US-only.

Remaining / honest granularity caveats:

4. Per-metro is not literally per-hex. Water and garbage are sourced at the per-municipality level and mapped onto the metro; every household in a metro carries that metro's figure, not a cell-specific bill. Within-metro variation in water and refuse rates is not resolved.
5. Some metros bundle garbage into taxes/rent → garbage shows $0. In this build that applies to Baltimore, Boston, Cincinnati, Houston, Kansas City, New York, Philadelphia, Pittsburgh, and Washington DC. Where refuse is folded into property taxes or rent, a separate garbage line would double-count, so it is held at $0; this is a coverage choice, not a true zero cost.
6. Cell, streaming, and broadband are national flat. $113.25 cell + $35 streaming + $89 broadband do not vary by metro; the connectivity-plus-streaming gap ($237.25/mo) is identical everywhere.
7. H+T burden ratios are still 2022 ACS. Only the U basket and the measured-T dollar columns are 2026. The H% and T% splits — and therefore the H+T% combined ratio and the threshold counts — are 2022. Vintages are labeled per-column; do not re-date the ratios.
8. 6 of 30 metros lack a measured precompute (Riverside, San Diego, Cleveland, Cincinnati, Kansas City, Baltimore) — their transport stays H+T Index modeled. Riverside also has no metro U row in the build and uses the US-national U total ($639.38/mo) as its fallback.
9. 22 of 24 measured metros have flat transit fares in this precompute (loaded GTFS lacks zone/OD rules) — non-NYC transit $ is a single-boarding floor, not distance-graded. Only NYC is OD/transfer-graded. Action: ingest official-feed zone/OD fares before publishing measured transit costs for non-NYC metros.
10. DC measured transit is thin (7.6% transit-cell coverage, 751 cells) — treat its measured transit median cautiously.
11. Measured T excludes car ownership. For a total-cost-of-ownership figure, add CNT's auto-ownership line; the report compares the commute slice (CNT running = gas/VMT + transit).
12. HUD income limits still sparse (CT/NY/NJ/PA, 319 rows); AMI band narrative elsewhere uses the H+T Index ACS-derived AMI. HUD's 80% limit runs ~+47% above CBSA median family income.

Prepared as a draft for review. Methodology and lineage credit: Center for Neighborhood Technology & Center for Housing Policy, "Losing Ground" (2012), and the CNT Housing + Transportation Affordability Index. All 2026 figures are our own, computed from the commute database and our proprietary multimodal cost algorithm.