The US dollar is losing 3.55% of its purchasing power per year.
We price one unchanging cart of everyday essentials — rent, groceries, gas, healthcare — and never swap in cheaper substitutes the way the official CPI does. That cart costs 147% more than in 2000, so a 2000 dollar buys about 41¢ of it today.
The Fixed Basket vs. CPI-U since 2000
Both rebased to January 2000 = 100. The Fixed Basket runs
1.00 pp/yr hotter than CPI-U
(3.55%/yr vs
2.55%/yr) — CPI adjusts for
substitution and quality; the fixed basket does not. Both series are
not seasonally adjusted, so each shows a real intra-year
cycle (energy-driven); rates are quoted year-over-year, which cancels it.
BLS CPI-U (CUUR0000SA0) is shown for comparison only — it is
not used anywhere in the index.
Annual rate over time
Year-over-year change in the index, %/yr — the same data as the chart above, read as a rate instead of a level (same-month differencing, so seasonality cancels). BLS CPI-U is shown for comparison only, not used in the index.
The lowest-income basket loses purchasing power 0.51 percentage points per year faster than the highest — 3.86%/yr vs 3.35%/yr — because it leans more heavily on healthcare and rent, the two fastest-rising categories. An unequal burden a single headline number can't show.
See the lowest-quintile basket →How it's built
Why not just use CPI? CPI assumes that when something gets pricey, you switch to a cheaper substitute — so it tracks the cost of keeping the same standard of living. Useful, but it measures coping with rising prices, not how much your money still buys. We measure the second thing: the cost of the exact same cart, year after year. See the full methodology.
Developers: the JSON API is live at
/api —
GET /api/v1/index/headline.
One short email when the numbers update. No spam, unsubscribe from any email.