Reported inflation and experienced inflation regularly diverge. The gap is explained by how the index is constructed rather than by any manipulation of the figures.
The index describes an average household
A price index weights each category by its share of average spending, so a rise in a category a particular household barely buys has little effect on their own costs.
Households differ enormously in composition. Renters, homeowners, drivers, families with young children and pensioners all face different effective baskets.
The published figure is accurate as an average and can be wrong for most individual households simultaneously, which is a property of averages rather than an error.
Frequently purchased goods dominate perception
People observe prices they encounter often — food, fuel, transport — and rarely observe prices for infrequent purchases such as appliances or insurance renewals.
Research on inflation perceptions has consistently found that frequently bought items carry disproportionate weight in what people believe prices are doing.
When those categories rise faster than the average, perceived inflation exceeds measured inflation even where the measurement is exactly right.
Quality adjustment is widely misunderstood
Statistical agencies adjust prices for changes in what a product delivers, so a device that costs the same while performing substantially better is recorded as cheaper.
The reasoning is sound, since an index should measure the cost of a constant standard of living rather than the price tag on a changing object.
Consumers experience the price tag, so the adjusted figure feels wrong even to people who accept the logic when it is explained.
Levels and rates are confused
A falling inflation rate means prices are rising more slowly, not that they are returning to earlier levels. Prices remain where the previous increases left them.
Households compare against remembered prices from several years ago, so the level gap persists in experience long after the rate has normalised.
This is the most common source of disagreement between official statistics and public sentiment during a disinflation.
Substitution assumptions matter
Index methods assume some substitution between goods as relative prices change, which reflects real behaviour but understates the cost for anyone who cannot substitute.
A household already buying the cheapest available option has no substitution left to make, and for them the unadjusted increase is the true one.
Statistical agencies publish variant indexes for exactly this reason, including measures for particular household types, though those figures rarely appear in coverage of the headline rate.