The simple version

The Bureau of Labor Statistics reported this week that consumer prices rose 3.4% over the past twelve months. That is the headline index, the one every news story quoted.

The same release published a second official inflation rate, the chained index, which came in at 3.3%. It is not a competing estimate or a correction. It is a different way of measuring the same thing, and since 2018 federal law has used it, rather than the headline number, to adjust income tax brackets and the standard deduction each year.

The numbers

• The Consumer Price Index for All Urban Consumers, the headline measure, rose 3.4% over the 12 months ending in July (Bureau of Labor Statistics, Consumer Price Index, July 2026, USDL-26-1378)
• The Chained Consumer Price Index for All Urban Consumers rose 3.3% over the same period, a tenth of a percentage point lower (same release)
• Federal law sets the annual adjustment from the chained index. The provision is Internal Revenue Code section 1(f), titled adjustments in tax tables so that inflation will not result in tax increases (26 U.S.C. 1(f))
• The statute compares the chained index for the preceding year against the price index for calendar year 2016 for the lower brackets, and 2017 for the higher ones (26 U.S.C. 1(f)(3))
• The chained index accounts for consumers substituting between goods as relative prices change; the headline index holds its basket fixed between periodic updates (Bureau of Labor Statistics)
• Measured December to December, the chained index has trailed the headline index by about 0.26 percentage points a year on average since 2018 (computed from the two published BLS series)
• It does not trail every year. In 2020 the chained index ran higher than the headline index, and in 2022 the two were within 0.01 of a point of each other (same computation)

Why one index runs slower than the other

The difference is substitution. The headline index prices a fixed basket: a set list of goods and services in set quantities, updated periodically. If beef gets expensive, the index keeps pricing the same amount of beef.

The chained index assumes people respond to prices. When beef gets expensive, some households buy chicken instead, and the chained index adjusts its basket continuously to reflect that shifting. Because substitution moves spending toward cheaper alternatives, the chained measure usually comes in a bit lower, which is what happened this month.

The economic argument for it is serious and straightforward: if people really do switch, a fixed basket overstates how much their actual cost of living rose. The argument against is equally straightforward: substituting chicken for beef is a real reduction in what you are consuming, and a measure that treats it as costless understates the squeeze. Economists have made both arguments for decades and this article does not settle them.

The published record cuts against the simple story in a way worth knowing. Comparing the two series December to December, the chained index has trailed the headline by about 0.26 percentage points a year on average since 2018. But in 2020 it ran higher, and in 2022 the two were effectively tied, because prices were moving on supply disruptions rather than on shoppers switching between goods.

What the tenth of a point does

Tax brackets are adjusted upward each year so that inflation alone does not push people into higher rates, a problem known as bracket creep. The statute that does this is titled, in its own words, adjustments in tax tables so that inflation will not result in tax increases. The size of that adjustment is set by the index the law names.

Because the chained index rises more slowly in most years, brackets and the standard deduction widen more slowly than they would under the headline measure. Any single year's difference is small. Repeated annually, it compounds, and thresholds drift lower relative to where the headline index would have put them.

Whether that is the right policy is genuinely contested and has been argued in Congress and among economists since well before the current law. What is not contested is the arithmetic: the index chosen determines the adjustment, and the two indexes give different answers.

The Real Cost lens on compounding a tenth of a point

Put the gap on a stated threshold so the size is visible. Every assumption is stated, and this is an illustration rather than anyone's tax outcome.

• This month the chained index came in 0.1 percentage point below the headline index, at 3.3% against 3.4%
• A threshold of $100,000 indexed at 3.4% a year for 20 years reaches about $195,200; indexed at 3.3% it reaches about $191,400
• That is a difference of roughly $3,700 on the threshold itself, about 1.9% lower after two decades
• This assumes the gap recurs every year at this month's size, which the published record shows it does not; the eight-year average gap is smaller, and in two of those years it ran the other way
• The mechanism runs the same direction for the standard deduction and other indexed parameters, all adjusted using the same index

How any of that affects a specific household depends on income, filing status, deductions, and circumstances this article cannot see, and those are questions for a credentialed tax professional rather than a news piece. The point here is narrower: a definitional choice made in one law, using an index most people have never heard of, adjusts a number on nearly every tax return every year.

What this means

When someone quotes an inflation rate, the useful follow-up is which index, because the government publishes several and they are wired to different things. The headline measure drives the news. The chained measure drives the tax code, and a third measure drives Social Security benefits, which we have covered separately.

It is a reminder of how much ordinary financial life runs through statistical definitions almost nobody reads. A methodology decision, made once, quietly adjusts a threshold on your tax return every year for the rest of your working life, and the only place it is visible is a monthly release that publishes both numbers side by side.

What this is NOT

This is not tax advice and it is not guidance on filing, planning, withholding, or any tax decision, all of which depend on individual circumstances and belong with a credentialed professional. This is not an estimate of anyone's tax bill, bracket, or liability. This is not a position on whether the chained index should be used for indexing the tax code, which is a real and long-running policy debate with serious arguments on more than one side that this article does not take. This is not a prediction of inflation, tax brackets, or future legislation. This is not advice about any security, fund, or asset. The compounding illustration assumes a gap that recurs at one month's size, which the published record shows is not what happens. This is not investment or financial advice of any kind.

Sources

• U.S. Bureau of Labor Statistics, Consumer Price Index, July 2026, USDL-26-1378: https://www.bls.gov/news.release/archives/cpi_08122026.htm
• U.S. Bureau of Labor Statistics, Chained CPI overview: https://www.bls.gov/cpi/additional-resources/chained-cpi.htm
• 26 U.S.C. 1(f), adjustments in tax tables so that inflation will not result in tax increases: https://uscode.house.gov/view.xhtml?req=granuleid:USC-prelim-title26-section1&num=0&edition=prelim
• Internal Revenue Service, annual inflation adjustments: https://www.irs.gov/