Hedonic adjustment is the methodology that price statisticians use to separate “the price of the product” from “the price of the qualities embedded in the product.” A 2010 laptop and a 2026 laptop are nominally the same item in the CPI basket: a personal-computer category. But the 2026 laptop has 16x more memory, a substantially better display, and a battery that lasts 3x longer. The quality has changed. If we treat the two laptops as identical, we miss the implicit price reduction.
Hedonic regression is the BLS methodology for handling this. It is technically valid and practically important; it is also the source of recurring public confusion about why CPI “underreports” inflation in technology categories.
The regression
For a category like laptops, BLS analysts assemble a panel of products with full specifications: CPU model, RAM, storage capacity, screen resolution, battery life, weight, and so on. A regression of the price against the specifications produces coefficients that can be read as “the implicit market price of an extra gigabyte of RAM” or “the implicit market price of an additional hour of battery life.”
When a model is replaced with a successor that has more RAM at the same shelf price, the hedonic adjustment computes how much of the “same” price represents a quality increase, and reports an effective price reduction in the index. The methodology is published in the BLS CPI Handbook; the equivalent ONS approach is in the Eurostat Manual on CPI Methodology.
Where it has the largest effect
- Personal computers and consumer electronics. Hedonic-adjusted “laptop CPI” has fallen approximately 90 % since 1998. The shelf price of a competent laptop in nominal dollars is roughly stable; the implicit per-spec price has collapsed.
- Televisions. Same dynamic, even more dramatic: hedonic-adjusted TV prices have fallen ~95% since 2000.
- Mobile phones. Phone hardware has improved continuously while phone shelf prices have fallen modestly. Hedonic adjustment captures the difference.
- Cars (partial). Quality features added over decades (anti-lock brakes, airbags, infotainment) are partially hedonic-adjusted but the methodology is more contested for cars than for electronics.
The criticism
Reference: hedonic-adjusted CPI for selected categories (US, 1998 = 100)
| Category | 1998 | 2010 | 2020 | 2026 |
|---|---|---|---|---|
| Personal computers | 100 | 31 | 14 | 10 |
| Televisions | 100 | 29 | 10 | 7 |
| Audio equipment | 100 | 52 | 28 | 19 |
| Photographic equipment | 100 | 62 | 38 | 29 |
| Apparel (no hedonic) | 100 | 119 | 122 | 130 |
| Food (no hedonic) | 100 | 131 | 156 | 192 |
The first four categories are heavily hedonic-adjusted. The last two are not. The divergence is striking and intentional: it reflects the methodology's view that electronics genuinely deliver more product-per-dollar over time, while a loaf of bread delivers approximately the same product across decades.
What this means for the inflation calculator
The published headline CPI already incorporates hedonic adjustment. When you use the calculator on this site, you are using a hedonic-adjusted figure whether you know it or not. For most everyday purposes — comparing a salary across decades, computing inflation-equivalent contract values — this is correct and consistent with what official statistical agencies publish. For technology-specific or quality-aware comparisons (e.g., “what would the 2026 equivalent of a 1998 desktop computer cost?”), the hedonic-adjusted figure may understate the apparent price drop you would actually observe at retail.