The Homes Most Likely to Need an Upgrade Are the Least Able to Pay for One

Two studies that could not say who lives in the stock they counted

HyreElectrical’s hazard-era panel study counted 67,068,006 US housing units built 1950–1989, the years Federal Pacific and Zinsco load centers were on the market.

Its follow-up on tenure found that 21,710,304 of the occupied ones are rented, and that a tenant in one of them has no legal standing to order the panel replaced.

Neither study could say whether the household actually living in that stock is one already struggling to pay for energy at all — Census ACS, the source for both, does not ask.

The Department of Energy’s own electrification research states the constraint plainly: "upgrading panels is not always feasible — especially for low-income families." That is a program-level observation, not a measurement. HYRE analysis.

The one federal instrument built to measure it directly is RECS, which interviews households about their own financial experience with energy costs rather than inferring it from housing age.

Cross-tabbing its energy-insecurity questions against the characteristics this site’s earlier studies already established — construction era and, by extension, tenure — is the natural next cut, and nobody appears to have published it against this specific hazard-era framing before.

The construction-era cross-tab: real, and not where you might expect it

Summing RECS’s four year-built rows inside the 1950–1989 window: 17.62 million of the 59.88 million occupied homes in that era report at least one form of energy insecurity — 29.4 percent.

The relative standard error is 1.9 percent on the summed estimate, computed as the root sum of squares of the four component RSEs, the same method this site’s ACS studies use for a derived sum.

Outside that window — everything built before 1950 or after 1989 — the rate is 15.96 million of 63.65 million, or 25.1 percent.

HYRE analysis, and the honest complication. The gap is real but it is not the sharpest cut available, and it is not unique to the specific years this site’s panel-hazard study cares about.

Homes built before 1950 — outside the FPE/Zinsco installation era entirely — report energy insecurity at 29.4 percent, statistically indistinguishable from the 1950–1989 window itself.

What actually separates the data is a line at 1990: every decade band built earlier runs 27.7 to 30.7 percent, and every band from 1990 onward runs 21.1 to 24.2 percent.

A page that reported only the 1950–1989 number against "everything else" would flatten this pattern; publishing all nine bands is the only honest way to show it.

The more likely driver: insulation, not age

The same table cross-tabs energy insecurity against a home’s insulation adequacy, self-reported by the household, and the gap there is far wider than the one by construction era.

Well-insulated homes report energy insecurity at 19.2% (6.58 million of 34.34 million); poorly insulated homes at 46.0% (9.80 million of 21.29 million); homes with no insulation at 51.0% (1.85 million of 3.63 million). That is a wider spread than the construction-era gap and almost as wide as the income gradient below.

HYRE analysis. Older homes are more likely to be poorly insulated, so construction era is plausibly standing in for insulation and equipment condition rather than measuring an independent effect of its own — and certainly rather than measuring anything about a panel specifically.

RECS’s public tables do not cross insulation against year built either, so this cannot be resolved into "how much of the era effect is really an insulation effect" from the tables available; it is named here as the more proximate and more likely explanation, not proven as one.

The income cross-tab: an eight-times gradient, unbroken

RECS asks 2020 annual household income in eight brackets, and the energy-insecurity rate falls in an unbroken line across every one of them.

It runs from 58.0 percent of households earning under $5,000 a year, 56.1 percent at $5,000–$9,999, 46.8 percent at $10,000–$19,999, 39.7 percent at $20,000–$39,999, 29.3 percent at $40,000–$59,999.

Above that it is 20.1 percent at $60,000–$99,999, 11.1 percent at $100,000–$149,999, and 7.2 percent at $150,000 or more.

HYRE analysis. This is the sharpest and cleanest gradient on this page — an eight-times difference between the bottom and top brackets, with no reversal at any step.

It is also, unlike the construction-era finding, exactly the shape a household-budget explanation predicts: energy insecurity as EIA defines it is fundamentally about the ability to pay, and income is the most direct measure of that ability the survey collects.

The tenure cross-tab: renters, roughly double, on every measure

Owner-occupied households report energy insecurity at 20.4 percent (16.92 million of 82.92 million); renter-occupied households at 41.0 percent (16.66 million of 40.61 million) — almost exactly double, from a smaller base of homes.

The gap holds on every one of EIA’s six specific measures: renters forgo food or medicine to pay an energy bill at 31.6 percent against owners’ 14.2 percent, and receive a disconnect or delivery-stop notice at 16.9 percent against 6.6 percent.

HYRE analysis. This site’s renter-tenure study already found that 48.7 percent of renter-occupied homes sit in the 1950–1989 hazard-era window, against 46.2 percent of owner-occupied homes — a real but modest gap in the same direction as this page’s construction-era finding.

Laid on top of a roughly two-times energy-insecurity gap and this site’s established fact that a tenant cannot lawfully authorize a panel replacement, the three findings point the same way without any one of them, alone, proving the others.

The household most likely to be renting a hazard-era home is also, independently, the household EIA finds most likely to already be behind on the energy bill — and the one with no legal standing to fix the panel even if it could afford to.

The one geography RECS actually supports: census region

RECS’s sample supports national, four-region and nine-division estimates — nothing finer. The South, which reports the largest volume of energy-insecure households at 14.00 million, also reports the highest rate: 29.9 percent of its 46.84 million occupied homes, against 26.9 percent in the West, 25.3 percent in the Northeast and 24.3 percent in the Midwest.

A scope note, stated because this site works from a five-state contractor store that is 74 percent Florida.

The South region in this table is Census-defined and includes Florida along with sixteen other states and DC; nothing in this section, or anywhere else on this page, uses HyreElectrical’s own contractor records.

The South’s highest-in-the-country rate is federal survey data, reported here because it is true, not because it happens to overlap with where this site’s own coverage is concentrated.

Method

Primary source. US Energy Information Administration, 2020 Residential Energy Consumption Survey, Table HC11.1, "Household energy insecurity."

The workbook states a final data release date of August 2025, revising a March 2022 preliminary release "to correct energy bill payment method rows" in a May 2022 update; the figures on this page come from the August 2025 final release, retrieved 6 September 2026.

Derivation. The 1950–1989 sums are the four RECS year-built rows "1950 to 1959," "1960 to 1969," "1970 to 1979" and "1980 to 1989" added together, for total occupied units and for each of four outcome columns (any insecurity, forgo food or medicine, unhealthy temperature, disconnect notice).

Relative standard errors on the summed "any insecurity" estimate are combined as the square root of the sum of the squared component absolute errors (each row’s published RSE multiplied by its value), the same root-sum-of-squares method this site’s ACS-based studies use for derived sums.

It is applied here to EIA’s published RSEs rather than the Census Bureau’s margins of error, which are not the same statistic: EIA states its RSEs at one standard error, not the Bureau’s 90 percent confidence convention.

What this page reuses, and what it does not. The 1950–1989 window and its 67.1-million-unit ACS total are established in the hazard-era study and are cited, not re-derived. Every RECS figure on this page — the year-built, income, tenure and insulation cross-tabs — is newly pulled from HC11.1, a table neither earlier study uses.

Reproducibility. HC11.1 is a single public workbook with a companion RSE sheet. Anyone with the file can rebuild every sum and percentage on this page from the rows cited.