CPI-W vs. CPI-E: The Index Debate Behind Every COLA
Side by side
| CPI-W (current law) | CPI-E (proposed) | |
|---|---|---|
| Population measured | Urban wage earners and clerical workers — about 29% of the U.S. population | Households with a reference person or spouse aged 62+ — about a quarter of the population |
| Medical care weight | Reflects working households' smaller healthcare share | Roughly double the CPI-W share — the single biggest difference |
| Housing weight | Large | Larger still — older households spend proportionally more on shelter |
| Transportation, food, education | Larger shares (commuting, tuition-age families) | Smaller shares |
| Status | Statutory escalator index since 1975 (§215(i)) | Experimental BLS research series; computed back to December 1982 |
| Long-run pace | — | ≈ +0.2 points per year faster than CPI-W (BLS characterization) |
A methodological caveat the BLS itself raises: the CPI-E is built by reweighting the same price quotes collected for the other indexes — it does not survey prices at senior-specific outlets or for senior-specific products, and its sample is smaller. It measures a retiree budget against general prices, not retiree prices as such.
The positions
For switching to the CPI-E: seniors' organizations, including AARP and The Senior Citizens League, argue the COLA should measure the inflation beneficiaries actually experience, and that the 0.2-point annual gap is exactly the slow purchasing-power leak their studies document (see purchasing power). Bills to make the switch are introduced in most Congresses.
Against, or cautious: the cost. A permanently larger COLA compounds across every beneficiary and every year; the SSA's actuaries have consistently scored a CPI-E switch as widening the program's long-range shortfall, which opponents say must be paid for explicitly rather than adopted as a technical correction. Methodologists add the BLS's own experimental designation as a reason to improve the index before legislating it.
The opposite proposal: the chained CPI-U, which grows more slowly, has been advanced in budget negotiations as a savings measure — and opposed by the same seniors' organizations for the mirror-image reason. The two proposals frame the real choice: the index is not a neutral measuring stick but a policy lever worth roughly two-tenths of a point a year, in one direction or the other, forever.
Current-law status
As of the 2027 cycle: the CPI-W remains the statutory index, no switch in either direction has been enacted, and the 2028 COLA will also be computed from third-quarter CPI-W. Any change would arrive as legislation and would be logged on the COLA news page and the legislative timeline.
Frequently asked questions
What is the difference between the CPI-W and the CPI-E?
Both are Bureau of Labor Statistics price indexes. The CPI-W tracks the spending of urban wage earners and clerical workers (about 29% of the population) and is the index the COLA law uses. The CPI-E tracks households aged 62 and over, whose budgets weight medical care and housing more heavily and transportation, food and education less.
Which index rises faster?
Over long spans, the CPI-E — by about 0.2 percentage points a year since the early 1980s, in the BLS's own characterization. The gap compounds: over a 25-year retirement it is the difference between benefits keeping pace with the official index and keeping pace with a retiree-weighted one.
Why doesn't Social Security just use the CPI-E?
Three reasons recur in the debate: the statute (§215(i)) names the CPI-W, so a change requires Congress; the CPI-E is designated experimental — the BLS builds it by reweighting the same price quotes rather than surveying senior-specific prices and outlets, on a smaller sample; and the SSA's actuaries score a switch as increasing the program's long-range cost, since every future COLA would be slightly larger.
What is chained CPI, and how is it different?
The chained CPI-U is a third index that accounts for consumers substituting toward cheaper goods as prices change; it typically rises about 0.25–0.3 points a year more slowly than the regular CPI-U. It has been proposed (notably in 2013 budget negotiations) as a way to slow benefit growth — the opposite direction from the CPI-E. Neither switch has been enacted; current law remains CPI-W.
Sources: Bureau of Labor Statistics, CPI-W and CPI-E program documentation (population coverage, experimental status, long-run comparison); Social Security Act §215(i); SSA Office of the Chief Actuary scoring practice for index-switch proposals.