Social Security COLA calculator.

The Social Security cost-of-living adjustment is one of the few automatic federal indexation programs that has continued largely unchanged since its 1975 introduction. The calculation, defined in statute, runs as follows:

1. Take the CPI-W (not CPI-U) for July, August, and September of the current year. Average them.
2. Take the same three-month average from the last year in which a COLA was paid.
3. Calculate the percent increase. If positive, that becomes the COLA, rounded to the nearest tenth of a percent. If zero or negative, there is no COLA (it cannot be negative).
4. Apply the COLA to all Social Security benefits effective with the December payment.

The SSA announces the upcoming COLA in mid-October, typically the same day the September CPI report is released. Beneficiaries see the increase in their first January payment.

Why CPI-W and not CPI-U?

When automatic COLAs were enacted in 1972 (taking effect 1975), CPI-W was the only urban CPI BLS published. CPI-U was introduced two years later. The statutory tie to CPI-W remained, and Congress has not modified it despite multiple proposals to switch to CPI-U, C-CPI-U, or the experimental CPI-E (which tracks spending patterns of households headed by someone 62 or older).

The three indexes typically move within a few tenths of each other, so the choice of index makes only a small difference in any single year. But across decades, the cumulative impact can be meaningful — a switch to C-CPI-U, for example, has been estimated to reduce average annual COLAs by roughly 0.3 percentage points.

When there's no COLA

Three years since 1975 have produced no COLA: 2010, 2011, and 2016. In each case, CPI-W in the Q3 average was below the prior Q3 average. By statute, a falling price level cannot produce a benefit cut — benefits hold steady until the index recovers to its prior high. The 2010–2011 episode followed the 2008 oil collapse and was politically contentious despite being mechanically correct.

The COLA is designed to preserve nominal benefit purchasing power against CPI-W. It does that on average across the basket, but several factors mean any individual retiree's real benefit can drift even with COLA in place.

Personal inflation differs from CPI-W. Retiree spending is weighted more heavily toward healthcare and housing than CPI-W reflects. The experimental CPI-E typically runs 0.2 to 0.3 percentage points higher per year than CPI-W. Over 20 years, that gap compounds to a roughly 5% shortfall in real purchasing power.

Medicare Part B premiums are deducted from benefits. When Part B premiums rise faster than COLA, net benefits can decline in real terms even when the gross COLA looks healthy. The "hold harmless" provision protects most beneficiaries from net decreases, but the protection isn't universal.

Taxation thresholds aren't indexed. The income thresholds that determine whether Social Security benefits are taxable have not been adjusted for inflation since the 1980s. As nominal benefits rise with COLA, more beneficiaries cross into taxable territory each year.

For deeper context on the underlying inflation measure, see our CPI report and categories page.