Every month, a single number moves markets, adjusts paychecks, and decides how much retirees receive from Social Security. That number is the Consumer Price Index, or CPI, published by the U.S. Bureau of Labor Statistics (BLS). Over the year ending in May 2026, the all items CPI rose 4.2%, with energy prices up 23.5% and shelter costs up 3.4%. Plenty of households wonder why their grocery and rent bills feel different from what gets reported, but that gap is not a mystery once you understand how the number is built.
The CPI measures the average change in prices urban consumers pay for a fixed basket of goods and services, covering food, housing, medical care, and transportation. That basket sounds simple, but the formula behind it has been revised repeatedly since the 1980s, and those revisions are the source of most public confusion about the figure.
The most consequential review began with the Advisory Commission to Study the Consumer Price Index, known as the Boskin Commission, convened by the Senate Finance Committee in 1995. Its final report, issued in December 1996, concluded the CPI overstated the true cost of living by roughly 1.1 percentage points per year, largely due to substitution bias: when one good’s price rises, consumers shift to cheaper alternatives, such as chicken instead of beef, which a fixed basket misses. A later GAO review found that between the Boskin report and mid 1999, the BLS made seven methodological changes tied to its findings.
One major change arrived in January 1999, when the BLS began using a geometric mean formula for most basic indexes instead of the older arithmetic average. Under the old approach, a price increase flowed straight into the index with no allowance for consumer behavior. The geometric mean formula assumes shoppers shift modestly toward close substitutes when one item in a narrow category gets pricier, such as switching brands of canned tomatoes, so it dampens the measured increase accordingly. In 2002, the agency added a second tool, the Chained Consumer Price Index, which works at a broader level: rather than assuming a fixed spending mix across whole categories like food or transportation, it recalculates the basket using updated spending patterns, capturing cases where consumers swap entire categories, such as eating out less and cooking more, as relative prices shift.
A separate source of skepticism is hedonic quality adjustment. When a product improves, such as a sharper television or faster phone, part of any price increase reflects that improvement, so the BLS uses regression models to estimate the quality portion and strips it out. A regression model is a statistical tool that looks at many products at once and isolates how much of the price difference between them is explained by each individual feature, such as screen resolution or processor speed, separating that quality related value from a true increase in price. This applies to a narrow set of categories, including computers, televisions, apparel, and some telecommunications services, and the agency says the net effect on the all items number is close to zero.
Critics argue these methods, taken together, understate the true cost of maintaining a household’s standard of living. The most prominent unofficial alternative came from Shadowstats.com, run by economist Walter “John” Williams, who recalculated inflation using pre 1990s methodology and produced figures consistently several points above the official CPI. Mainstream economists have largely been skeptical: Niskanen Center fellow Ed Dolan found Williams’ numbers implausibly high against interest rate and output data, and BLS economists published research addressing what they viewed as misconceptions in his approach.
The CPI also determines annual cost of living adjustments for Social Security, federal retiree, and food assistance programs, and factors into many private wage contracts and pensions, so a lower reported rate means smaller automatic benefit increases. None of this means the BLS acts in bad faith: the agency documents its methods publicly, and independent reviews, including the GAO’s, generally found the changes grounded in standard economic reasoning rather than political manipulation. Still, the CPI remains, as the BLS has acknowledged, an approximation of a cost of living index, not a perfect mirror of any one household. Knowing the mechanics behind the number makes it easier to ask: how has your own cost of living changed, and does the official figure capture it.
