The Social Security COLA Tracks a Basket Retirees Don’t Buy

The index excludes the jobless, the self-employed, part-time workers, and anyone who has left the labor force entirely: the very people who receive the adjustment. The number on the notice is the exact number the law demands, and that is precisely why it can feel so far off.

The Formula Runs on Its Own

The Social Security Administration does not decide what the COLA will be. Under the 1972 Social Security Amendments, Congress tied the annual increase to an automatic formula, and the first automatic COLA was paid in 1975. The calculation uses the percentage increase in the average CPI-W from the third quarter of the last year a COLA was determined to the third quarter of the current year. That means the Bureau of Labor Statistics releases the September CPI-W by mid-October, SSA plugs it into the statutory comparison, and the number is set. There is no hearing, no vote, no discretion.

Two hard rules follow from that design. The first: the increase equals the rise in the CPI-W third-quarter average, not any other inflation gauge. A consumer might see gas prices jump in February or a Medicare premium spike in January, but those months play no role. Only July, August and September matter. The second: if the average for that third quarter does not exceed the prior computation quarter, there can be no COLA. The law does not allow a small increase, a round-up, or a make-up check. Zero is zero.

Who Sits in the CPI-W Basket

The CPI-W represents households where at least one member has worked 37 weeks or more in the previous 12 months in an eligible occupation, and where half or more of the household’s income comes from wages tied to that work. Eligible occupations are blue-collar and clerical jobs; professional and salaried workers are out. So are the part-time, the self-employed, the unemployed, and households in which nobody is in the labor force—which is to say, retirees. The Bureau of Labor Statistics calls it a subset of the broader CPI-U, covering roughly 30 percent of the United States population. Its original purpose was to escalate wages and benefits in collective-bargaining agreements for blue-collar workers, and the BLS still describes it as used primarily for that kind of adjustment.

A retired couple living on Social Security does not meet any of those criteria. Yet the formula that sets their annual COLA looks exclusively through the lens of a household in which someone is clocking in at a factory or an office every week. That is not an oversight; it is the index Congress wrote into the statute.

Why It Feels Like a Shortfall

Retirees’ budgets tilt heavily toward medical care and shelter. A working household of clerical earners spends a different mix: more on transportation to get to a job, more on clothing for work, different food-at-home patterns. The CPI-W’s weights reflect that working-age consumption, not the spending of older, non-working Americans. Since the index population explicitly excludes the retired, the price changes it captures do not track what a fixed-income 72-year-old experiences at the pharmacy counter or the dentist’s office.

That disconnect does not mean the COLA is wrong. It means it is answering a question that is not “how much more did retirees pay?” The math is faithful to the CPI-W third-quarter average. If the index says medical care rose 2 percent but fuel oil jumped 10 percent, the COLA will reflect the full basket, and a retiree who drives rarely but sees a cardiologist frequently will feel the gap. The BLS itself notes that the index is designed for blue-collar cost-of-living adjustments. Nobody has ever claimed it was built for an aging, non-working population.

The Years the COLA Disappeared

The zero-COLA rule is not theoretical. Since automatic annual adjustments began in 1975, three years have produced no increase payable: 2010, 2011, and 2016. Each time, the average CPI-W for the relevant third quarter did not exceed the prior cost-of-living computation quarter. Energy prices can tumble, a recession can flatten broader inflation, and the formula simply delivers a zero. For the beneficiary, that means the same deposit every month while the world outside feels no cheaper. Congress did not intervene, and the statute provided no alternative route. The calculation ran, the number came back flat, and SSA’s obligation was discharged.

The 2026 Increase, Step by Step

The most recent application of the formula appeared when the Social Security Administration published its 2026 COLA fact sheet. The increase is 2.8 percent, based on the rise in the CPI-W from the third quarter of 2024 through the third quarter of 2025. That is the statutory third-quarter window, nothing more. BLS released the September 2025 CPI-W by mid-October, SSA pulled the average, compared it to the 2024 third-quarter average that had produced the previous COLA, and the difference was 2.8 percent. All beneficiaries and Supplemental Security Income recipients receive that percentage increase automatically.

A reader can reproduce it anytime. Take the CPI-W index levels for July, August and September of the current year, average them, and compare with the same three-month average from the year in which a COLA was last triggered. The percentage change, rounded to the nearest tenth, is the COLA. That is the whole machine. It has no dial for whether nurses’ aides, home heating oil, or Medicare Part B premiums rose faster. The only number that can come out is the one the CPI-W third-quarter average gives.

Fifty years after the automatic COLA began, the statutory architecture remains unchanged. The calculation works exactly as intended. The index, however, describes a household that stopped looking like a typical Social Security beneficiary the moment the last paycheck was received. The COLA is precise, repeatable, and locked into law. It just does not measure the basket that retirees actually buy.