Index methodology

What the number is, and is not

The published value is a composite index. A result of 113 means the weighted measures in this registry sit 13% above the national value of the same measures. It does not mean a household spends 13% more.

• It is not a statement that a typical household here spends this much more or less than a typical United States household.
• It is not a household budget, a market basket, or a spending forecast.
• It is an index of the measures listed below, weighted by category, relative to the national value of those same measures.

Known limits of the composite

These are properties of the method, not data gaps. They apply to every location.

• Within a category, every available measure carries equal influence unless explicit measure weights are recorded for the version.
• Energy averages electricity and natural gas equally, although natural gas use varies sharply by climate, heating system, and whether a home is connected to gas at all.
• A category represented by a single loaded measure gives that measure the category's entire weight. Transportation currently rests on commute time alone and is a lower-confidence category.
• Statutory facts that reduce cost, such as a state with no individual income tax, appear as a measure at or near zero rather than as a separate offsetting credit.
• Food, health care, insurance, goods, and child care have no weighted measure, so they cannot move the index in either direction.
• Contributing measures come from different reporting periods: ACS 5-year estimates, EIA energy releases, and state tax-year records are combined into one number.
• Tax measures built on externally sourced local averages are held at pending verification and excluded from the index until the publisher is confirmed.
• Some conceptual overlap remains between Housing and Energy: ACS gross rent includes applicable utility and fuel costs, while the Energy category measures statewide energy price pressure.

Eligibility of a measure

A measure can enter the calculation only when it has a verified published source, a known reporting period, a known geographic level, a known unit, a compatible national benchmark, a documented cost direction, approval for score version v1.0, and no pending verification or demonstration flag. Missing or invalid measures are never assigned a value of zero: they are excluded and reported as not available.

Indexing a single measure

Every cost measure is converted to an index where 100 equals the national benchmark: location value divided by benchmark, times 100. An electricity price of 18 cents against a benchmark of 15 cents indexes at 120, meaning that measure sits 20% above its national benchmark. Income is never treated as a cost and is never inverted on its own; it is handled by the purchasing-power step below.

Category weights

A category index is the mean of its available measure indexes. The overall index is the weighted mean of the available categories. When a category has no data for a location, its weight is removed and the remaining weights are renormalized, and the resulting coverage ratio is shown on the scorecard.

Context-only categories carry no weight until a verified national dataset is loaded for them: Goods and Services, Homeowners Insurance, Health Care.

Weighted and context-only measures

Only weighted measures move the index. Context-only measures are published beside it so a reader can see the fuller picture without them influencing the number.

• Housing (35%)

  Weighted: Median gross rent, Median home value, Housing-cost burden

  Context only: Selected monthly owner costs, Median rent as a percentage of income, HUD Fair Market Rent

• Energy and Utilities (15%)

  Weighted: Residential electricity price, Residential natural gas price

  Context only: Average monthly residential consumption, Average monthly electricity cost, Electricity cost at a selected consumption level

• Taxes (20%)

  Weighted: A single sales-tax measure, Effective property-tax proxy

  Context only: State sales tax component, Average local sales tax component, Top statutory individual income-tax rate, Tax structure and military retirement treatment

• Transportation (10%)

  Weighted: Mean commute time, as a pressure proxy

  Context only: Households without a vehicle, Vehicle availability, Mode of travel, Work from home

• Household Income and Purchasing Power (20%)

  Weighted: Purchasing-power pressure index

  Context only: Per capita income (median household income is the supporting input)

Avoiding double counting

A measure may not carry weight when it materially duplicates another weighted measure. These pairings are enforced automatically, and the index is marked provisional if the registry ever violates one.

• The combined sales tax rate already contains the state rate and the average local rate, so at most one of the three carries weight.
• The average monthly electricity bill is derived from the same price, sales, revenue, and customer records as the electricity price, so only the price is weighted.
• Selected monthly owner costs bundle mortgage payments, real-estate taxes, insurance, utilities, and fuels, so they are context only while the property-tax proxy and energy prices are weighted.
• Median rent as a percentage of income overlaps the housing-cost burden share, so only the burden share is weighted.
• Per capita income duplicates median household income, so it is context only.

Income and purchasing power

Income is not a cost, so a higher income is never reported as a higher cost. Median household income is indexed against the national median. Housing, Energy, Taxes, and Transportation are combined with their weights normalized among themselves to form a non-income cost index. Purchasing-power pressure is the non-income cost index divided by the income index, times 100. That single figure is what the 20% income and purchasing-power weight carries, which keeps the calculation from feeding back into itself. Per capita income stays context only so income is not counted twice.

Coverage, applied weights, and the publication threshold

Coverage is the share of the original intended category weight that had eligible data, calculated before renormalization. If Taxes are unavailable, coverage is 80%, not 100%. The remaining weights are then renormalized into applied calculation weights: Housing 43.75%, Energy 18.75%, Transportation 12.5%, and income and purchasing power 25%. Both figures are shown whenever a category is missing. An overall index is published only when coverage is at least 60%, Housing is available, and every included measure has a compatible benchmark. Otherwise the index reads as not available with the blocking reason stated.

The five-level scale

Index values are grouped into five levels. Each level has a color and a text marker, so the scale is readable without relying on color.

• − −Much lower costIndex below 92
• −Lower costIndex 92 to 97.5
• =Near benchmarkIndex 97.5 to 102.5
• +Higher costIndex 102.5 to 110
• + +Much higher costIndex above 110

Geographic resolution and specificity

Census measures are reported for states and counties. Energy prices are reported by state, so a county view reuses its statewide value and labels it as a statewide value displayed for county context. Nothing is interpolated between geographies, and no value is invented where an agency reports none.

Coverage and geographic specificity are separate. Specificity is the share of contributing measures actually reported at the selected geography. A county can have 100% data coverage and 72% county specificity when some inputs are statewide.

Reporting periods

Each measure keeps its own reporting period, and the score shows the oldest and newest period among its inputs, for example score inputs span 2023 to 2025. A single year is never assigned to the composite unless all material inputs share it.

Margins of error

ACS estimates are survey estimates and carry margins of error, which matter most for small counties. Where the published margins are stored for a measure, they are shown with it, and a small difference between two places may not be statistically meaningful given those margins.

Reading the number

A result of 113 is stated as: the CostExplore Cost Pressure Index is 113, compared with the national benchmark of 100, or measured cost pressure is 13 index points above the national benchmark. It is never stated as a household spending 13% more.

Versioning

Measure membership for v1.0 is frozen. The exact list of weighted and context-only measures is recorded against the version, so a dataset added to the system later is context only for this version and cannot change a v1.0 score even though the displayed category weights stay 35/15/20/10/20. Admitting a new measure to the weighted set requires a new score version.

Every calculated score stores its score version, category and measure weights, eligible and excluded measures, benchmark and source records, reporting periods, coverage ratio, geographic specificity, calculation timestamp, and calculation version. Changing category weights, weighted measures, cost direction, the purchasing-power formula, missing-data handling, or the interpretation thresholds requires a new score version with an effective date, a change log, a recalculation, and preservation of the prior version. The thresholds and weights on this page are CostExplore methodology choices, not weights published or endorsed by the Census Bureau, EIA, BEA, BLS, or another government agency.

Limitations

• Survey estimates carry sampling error, which is larger for small counties.
• Reporting periods differ between sources, so a comparison can mix a 2023 survey estimate with a 2025 energy price.
• The electricity estimator uses the EIA average residential retail electricity price, which EIA derives from retail revenues divided by retail sales and which includes delivered electricity costs such as generation, transmission, distribution, taxes, and fees. It is an average across a state's residential customers, not a utility tariff, so it does not reflect fixed monthly charges, tiered or time-of-use pricing, or an individual utility's rate structure.
• The score is a planning aid, not a household budget and not financial advice.