CI
TimePriceIndex.com

Methodology

Construction rules for the TPI-US 25, the flagship US time-price index. Versioned. Rules-based. No discretionary overrides.

Version

This document describes Methodology v1.0, effective at index inception. Changes to construction rules are recorded in the changelog on the About page and reflected in the version number on this page. Prior versions remain accessible by direct URL once superseded.

Definition

A time price is the quantity of work time required to earn the money to buy a unit of a good or service. It is the nominal price of the good divided by the nominal hourly wage. Expressed in minutes or hours, a time price restates affordability in units of life rather than units of currency.

The TPI-US 25 is a fixed-basket time-price index covering 25 US consumer goods and services. The index level is the time price of the basket in a given year, expressed as a percentage of its time price in the base year. A falling index value indicates rising abundance; a rising value indicates falling abundance.

Relationship to cost-of-living indices

The TPI-US 25 is a cost-of-living index in the broad sense: it measures the cost of acquiring a defined basket of consumer goods and services. It differs from price-based cost-of-living indices — including the US Consumer Price Index (CPI) and the Personal Consumption Expenditures price index (PCE) — in three structural ways.

First, the denominator is a wage series, not a deflator. CPI and PCE express the cost of the basket in current dollars and rely on a separately computed inflation series to compare across periods. TPI expresses the cost of the basket in hours of work at the prevailing average hourly wage, which removes the need for an inflation deflator.

Second, the basket is fixed for the life of a methodology version. CPI and PCE update their baskets periodically to reflect substitution. The TPI basket is held constant so that long-period comparisons describe the same set of goods.

Third, the items are equally weighted rather than expenditure-weighted. CPI and PCE weight by household consumption shares to track an aggregate cost of living. TPI treats each constituent as an independent affordability probe and reports their unweighted mean.

These choices produce a measurement adjacent to, not interchangeable with, CPI and PCE. The FAQ covers the comparison in further detail.

Base year

The base year is 1980. The index value in 1980 is set to 100. The basket is priced in nominal US dollars and divided by the 1980 nominal hourly wage to produce the base-year time price. All subsequent annual values are expressed relative to that base.

1980 is chosen for three reasons. It aligns with the base year used by the Simon Abundance Index and related work, supporting cross-comparison. It predates the consumer electronics revolution, which is the most striking source of time-price decline in the period. And it is the earliest year for which every constituent in the basket has a documented price observation from a primary source.

Wage denominator

The wage denominator is FRED series AHETPI, Average Hourly Earnings of Production and Nonsupervisory Employees, Total Private, published by the US Bureau of Labor Statistics. The annual value used is the simple average of the twelve monthly observations in that calendar year.

AHETPI is selected over alternatives for the following reasons:

  • Coverage of the base year. AHETPI is published continuously since January 1964, providing sixteen years of pre-base-year history and full coverage from 1980 forward.
  • Representative of typical earners. The series excludes managerial and supervisory pay, which has grown faster than the rest of the wage distribution. Including it would understate the time price faced by the median worker.
  • Transparent construction. AHETPI is computed from the Current Employment Statistics establishment survey using a documented methodology that has remained stable since inception.

Two alternatives were considered and rejected. CES0500000003, Average Hourly Earnings of All Employees, only begins in March 2006 and cannot be used for a 1980 base. Median weekly earnings from the Current Population Survey are reported weekly rather than hourly and would require an hours-worked assumption to convert.

Basket selection rules

An item qualifies for inclusion in the TPI-US 25 if and only if it satisfies all of the following:

  • Continuous price series. A primary-source nominal US price observation exists for every year from 1980 through the current annual reporting year, with no gaps longer than two consecutive years requiring interpolation.
  • Consumer relevance. The item is purchased directly by US households, not solely as an industrial input. Wholesale commodities and producer prices are excluded.
  • Defined unit. The item has a stable, specifiable unit of measurement (one gallon, one dozen, one kilowatt-hour, one month of service, one ticket). Quality changes over time are addressed in the quality adjustment rule below; unit definitions are not.
  • Independent source. No two items in the basket draw their entire price series from the same single subseries of a single agency, to limit source-correlated error.

The basket is constructed to span six categories: food, energy, housing, transportation, communication and electronics, and services. The category allocation is documented on the Constituents page. The category structure is informational; it does not enter the index calculation.

Index calculation

For each constituent i in year t:

  • Let Pi,t be the nominal US price of one unit of constituent i in year t, in dollars.
  • Let Wt be the annual average of AHETPI in year t, in dollars per hour.
  • The time price of constituent i in year t is Ti,t = Pi,t / Wt, expressed in hours.
  • The constituent's normalized index in year t is Ri,t = Ti,t / Ti,1980, expressed as a ratio relative to the constituent's own 1980 time price.

The basket index in year t is the unweighted arithmetic mean of the 25 normalized constituent indices, scaled to 100: It = (100 / 25) × Σ Ri,t. By construction I1980 = 100. A value below 100 indicates that the basket has become more affordable in work-time terms; a value above 100 indicates the reverse.

Each constituent is normalized to its own 1980 time price before averaging. Without normalization, items with large absolute time prices — a median home, a year of tuition — would dominate the arithmetic mean and the index would track those items rather than the basket as a whole.

Equal weighting is chosen over expenditure weighting deliberately. Expenditure weights would cause the index to track CPI-style consumption-weighted inflation, which is already well measured elsewhere. Equal weighting treats each item as an independent affordability probe and surfaces the dispersion of time-price changes across the basket.

Reporting period

The index is published annually. The reporting year t is finalized once all constituent price observations and the AHETPI annual average for t are available from primary sources, typically in the second quarter of year t+1. Provisional values are not published.

Quality adjustment

Two constituents in the basket — a representative personal computer and a representative mobile phone — describe goods whose physical characteristics have changed substantially over the index history. For these constituents the price series is anchored to the BLS hedonic-quality-adjusted CPI subindex for the corresponding category, applied to a 1980 reference price for a then-typical unit. This treatment is documented per-constituent on the Constituents page.

For all other constituents no quality adjustment is applied. A 1980 dozen eggs and a 2025 dozen eggs are treated as the same unit for index purposes. Where this assumption is contestable (for example, the energy efficiency of household electricity end-use), the contest is noted in the constituent's source notes rather than adjusted for in the index.

Missing observations

If a primary source publishes a single-year gap, the value is interpolated linearly between the two adjacent years and the interpolation is flagged in the published data file. Gaps of two consecutive years are interpolated only if a secondary source confirms the trend; gaps longer than two consecutive years disqualify the constituent and trigger a methodology revision.

Rebalancing and revisions

The basket is fixed at index inception and is not rebalanced on a schedule. A constituent is removed only if its primary source ceases publication or if the underlying good is no longer purchased by US households at a meaningful scale. A removal is replaced by the closest-fitting candidate satisfying the basket selection rules, the replacement is logged in the changelog, and the historical index is recomputed with the new constituent backfilled to 1980 to preserve the chain.

Primary-source price observations are sometimes revised after initial publication. The TPI-US 25 incorporates published revisions on the next annual reporting cycle. The full revised history is republished as a single CSV; no point-in-time snapshots are maintained.

Known limitations

The index is a transparent benchmark, not a measure of economic welfare. The following limitations are inherent to its construction:

  • Distribution. AHETPI is an average. Workers earning above or below the average face proportionally different time prices. The index does not characterize the distribution.
  • Tax and transfer. The wage denominator is gross hourly earnings before taxes, transfers, and benefits. Net take-home time prices differ.
  • Geography. All prices are US national averages. Regional time prices, particularly for housing, deviate substantially from the national figure.
  • Quality. Outside the two flagged constituents, no quality adjustment is applied. Where goods have improved in quality at a fixed unit definition, the index understates the abundance gain.
  • Basket scope. Twenty-five items cannot represent the full consumption bundle of a US household. The TPI-US 25 is a probe, not a census.
  • Hours worked. The index expresses prices in hourly-wage equivalents. It does not adjust for changes in average hours worked per year, which have declined modestly over the index history.

Citation

Suggested citation: Time Price Index, TPI-US 25, Methodology v1.0. TimePriceIndex.com. Underlying source series should additionally be cited per their issuing agency's guidelines.