The 2027 Social Security COLA Lands October 14: Why It Won't Feel Like a Raise

The raise is real. The reason it does not feel like one is also real, and it is mostly arithmetic.

Anyone who has sat at a kitchen table helping a parent open the annual Social Security letter knows the small deflation that follows. The percentage looks fine. Then January arrives, the deposit lands, and the number is not what the letter implied. Nobody made a mistake. Three separate mechanisms sit between the announced cost-of-living adjustment and the money that shows up in an account, and none of them are hidden — they are just never explained in the same place.

The Social Security Administration announces the 2027 adjustment on October 14, 2026. Early forecasts cluster in the mid-single digits, which would be a healthy headline number. Here is what happens to it on the way down.

The formula is narrower than most people assume

The COLA is not set by a committee and not indexed to the inflation figure you hear on the news. It is a mechanical calculation, fixed in law since 1975, and it uses a single input: the Consumer Price Index for Urban Wage Earners and Clerical Workers, known as CPI-W.

The arithmetic is simple enough to do yourself. Take the average CPI-W for July, August and September of this year. Compare it with the average CPI-W for July, August and September of the last year in which a COLA was payable. The percentage increase, rounded to one decimal place, is the COLA. If the figure comes out negative, the adjustment is zero — benefits are never cut for deflation.

Two consequences follow immediately, and both surprise people.

First, the measurement window is three months, not twelve. Whatever inflation did in February is irrelevant. A summer spike in gasoline prices raises the COLA; the identical spike in December does not touch it. This is not a flaw so much as a design choice with a side effect — the announcement in October is a snapshot of one quarter, dressed as an annual figure.

Second, the index is backward-looking by construction. The raise you receive in January 2027 is compensation for price increases that already happened, measured through September 2026. If prices accelerate over the winter, you carry that gap for a full year before the formula acknowledges it.

CPI-W measures a household that is not yours

Here is the part that deserves more attention than it gets. CPI-W tracks the spending of urban wage earners and clerical workers — people who are, by definition, working. It covers roughly a quarter of the U.S. population, and retirees are not the population it was built to describe.

The consequences show up in the weights. A working household spends more of its budget on transportation — commuting, vehicles, fuel — than a retired household typically does. A retired household spends considerably more on medical care and, in many cases, on housing-related costs it cannot easily reduce. Medical care has, over most of the last two decades, risen faster than the general price level.

The Bureau of Labor Statistics has published an experimental index for this exact problem since 1982: CPI-E, weighted for households headed by someone aged 62 or older. Over long stretches it has run modestly above CPI-W — a few tenths of a percentage point a year. That sounds trivial. Compounded across a twenty-five-year retirement, a few tenths a year is not trivial at all.

Proposals to switch the COLA to CPI-E surface in Congress regularly and have never passed. The reason is arithmetic too: a higher index means higher outlays from a trust fund already facing a scheduled shortfall. This is a genuine trade-off, not an oversight, and it is worth knowing that the index you are measured by was chosen partly for what it costs.

Medicare Part B gets paid before you do

For most beneficiaries, the Part B premium is deducted from the Social Security payment before the deposit is made. So the sequence in January is: apply the COLA, subtract the new Part B premium, deposit the remainder. The raise is real, but a share of it is spent before it becomes visible.

Part B premiums have generally risen faster than the COLA. When that happens in a year with a modest adjustment, a meaningful fraction of the increase — occasionally most of it — is absorbed on the way through.

There is a protection here that is genuinely useful and widely unknown. The hold harmless provision says that for most people who have Part B deducted from their Social Security benefit, the dollar increase in the Part B premium cannot exceed the dollar amount of their COLA. Your net deposit cannot fall from one year to the next because of a Part B increase. It can be flat. It cannot go backwards.

Three groups are outside that protection, and the exceptions matter:

People enrolling in Part B for the first time in that year.
People who pay their Part B premium directly rather than having it withheld from a benefit.
People paying the income-related monthly adjustment amount, IRMAA, because of higher reported income. Hold harmless does not apply to IRMAA surcharges, and IRMAA is assessed on a two-year lookback — your 2027 surcharge is determined by your 2025 tax return. A one-off event in 2025, such as selling a property, can raise a 2027 premium long after the money is gone.

The Part B figure for 2027 is announced separately, usually in November, a few weeks after the COLA. Until then, any net calculation is an estimate. Do the estimate anyway, then revise it.

What claiming early does, and does not, do to your COLA

A persistent worry among people who claimed at 62 is that they somehow receive smaller or fewer cost-of-living adjustments. That is not how it works, and the correction is worth stating plainly.

Cost-of-living adjustments begin accruing to your record in the year you turn 62, whether or not you have claimed. Someone who waits until 70 does not miss eight years of COLAs — they are applied to the primary insurance amount in the background and are already reflected in the benefit that starts at 70. Claiming early does not forfeit any adjustment.

But there is a real effect underneath the myth, and it is about compounding rather than eligibility. A COLA is a percentage applied to your benefit. Claiming at 62 permanently reduces that benefit — up to about 30 percent below the full retirement age amount, depending on your birth year. Every subsequent COLA is then a percentage of the smaller base, so the dollar gap between an early claimer and a late claimer widens every single year, even though both received the identical percentage.

A five percent adjustment on a $1,800 benefit is $90. The same five percent on a $2,600 benefit is $130. Next year, the gap compounds again. Over a long retirement that divergence is substantial. It is not a penalty added to early claiming — it is the original reduction, growing.

The historical "notch," incidentally, is a different thing entirely: a cohort born between 1917 and 1921 received lower benefits than adjacent cohorts because of how the 1977 amendments corrected a flawed indexing formula. It has nothing to do with claiming age, and it is now a matter of history rather than planning.

Your inflation rate is not the government's

The COLA measures a national average basket. You do not buy the national average basket. The gap between the two is the number that actually determines whether you are keeping up, and almost nobody calculates it.

It is not hard to approximate. Pick your five largest recurring expense categories — for most retired households some combination of housing, medical and prescription costs, food, transportation, and insurance. Compare what each one costs this year against last year, using actual statements rather than memory. Weight them by how much of your budget each represents.

What people usually discover is that one category is doing all the damage while the others behave. Homeowners insurance, a supplemental plan premium, or a property tax reassessment can move enough on its own to swamp a mid-single-digit COLA, while groceries and fuel sit quiet. A national average cannot see this, and neither can a headline.

A worksheet to net the raise down to reality

Run this in October once the COLA is announced, and again in November when the Part B premium is published. Ten minutes, on paper.

Current gross monthly benefit. Not the deposit — the gross figure, before Part B, from your most recent SSA statement or your account at ssa.gov.
Announced COLA percentage. Multiply line 1 by it. This is the gross raise.
New gross benefit. Line 1 plus line 2.
Current Part B premium. Including any IRMAA surcharge you are paying.
New Part B premium. Use the announced figure once available; before that, estimate using the last three years of increases.
Premium increase. Line 5 minus line 4. This comes straight out of line 2.
Net raise. Line 2 minus line 6. This is the actual change in your deposit.
Net raise as a percentage. Line 7 divided by your current net deposit. Compare it with line 2's headline percentage. The distance between those two numbers is the whole point of the exercise.
Your personal inflation rate. From the category exercise above.
The real answer. Line 8 minus line 9. Positive means you gained ground this year. Negative means you lost it, and by how much — which is the only figure here worth acting on.

Keep the sheet. Doing this for three consecutive years turns a set of annual surprises into a trend line, and a trend line is something you can plan against.

What the arithmetic leaves you

Very little of this is under your control, which is worth saying out loud rather than pretending otherwise. The index is set by statute. The premium is set nationally. The announcement lands when it lands.

What is in your control is narrow and real: whether you know your own inflation rate; whether you check that a Part B surcharge is based on a correct and current income picture — a life-changing event such as retirement itself can be appealed with form SSA-44; whether the one category driving your personal number is one you can actually renegotiate, which for insurance is more often than people assume.

And whether you look at line 10 at all, instead of the headline. A percentage announced in October is a fact about a national index. The figure on line 10 is a fact about your household, and it is the one that determines what next year feels like.

Questions worth asking

When exactly does the 2027 increase show up?

The adjustment applies to benefits for December 2026, which are paid in January 2027. Payment dates depend on your birth date, so the first larger deposit arrives on your usual January payment day rather than on the first of the month.

Can the COLA ever be zero?

Yes, and it has been three times in recent memory — 2010, 2011 and 2016 all had zero adjustments, following periods when CPI-W did not rise across the measurement window. Benefits are never reduced, but a flat year is entirely possible under the formula.

Is the COLA taxable?

The COLA is not taxed separately, but a larger benefit can push more of your Social Security into the taxable range, because the thresholds that determine how much is taxable — $25,000 and $34,000 for single filers, $32,000 and $44,000 for joint — are not indexed to inflation and have not moved since the 1980s. Each COLA therefore quietly increases the share of beneficiaries who owe tax on part of their benefit.

Why do the forecasts published before October keep changing?

Because only three months of data matter, and until September's CPI-W is published in October, part of the calculation is genuinely unknown. Forecasts made in spring are extrapolations from data that is not in the formula.

Should I delay claiming because of how COLAs compound?

The compounding described above is one input among several, alongside your health, other income, marital situation and whether you need the money now. It is a real consideration and not a decisive one on its own. This is a case for running your own numbers, ideally with someone who can see your whole picture.

What strikes me every year is the mismatch in precision. The formula is exact to a decimal place, applied to a basket that belongs to somebody else. Both things are true at once, and only one of them shows up in the letter.