There's hardly a statistic people trust less than the inflation rate. The agency publishes a figure and a large share of the population feels it has little to do with reality.

Interestingly, both sides have a point, and the reason lies in the methodology. It isn't manipulated, but it measures something other than what most people assume.

How the figure is produced

A consumer price index tracks a basket — a weighted collection of goods and services intended to represent the consumption of an average household.

The weights come from periodic surveys of spending behaviour. Housing is the largest block, then transport, then food, then leisure, and so on.

Price collectors gather a great many individual prices month after month, and the index is calculated from them. The inflation rate is the change in that index against the same month a year earlier.

Methodologically that's sound and internationally comparable. The problem arises elsewhere.

The average household doesn't exist

The central point: there is no average household. There are very different households whose spending structures diverge sharply.

A low-income household spends a considerably larger share on food, energy and rent — precisely the categories that have risen most in recent years.

A high-income household spends proportionally more on things whose prices rise slowly or even fall, consumer electronics being the obvious example.

The result is a real, measurable inequality in experienced inflation. Statistical agencies and research institutes have examined this, and the differences between income groups during periods of high energy prices are substantial.

So when somebody says inflation feels much higher, for many people that's simply accurate.

Perception is skewed

On top of that sits a psychological effect that's well documented.

We remember the prices of things we buy often. Bread, milk, coffee, fuel. We see those prices several times a week.

Rent is paid once a month by standing order. Insurance once a year. A television is bought every eight years, and the fact that it has become cheaper goes entirely unnoticed.

So frequently purchased goods dominate perception regardless of their actual share of spending. And price increases stick in memory more firmly than decreases — a well-established pattern in behavioural economics.

Quality adjustment

The most contested methodological point, and the one critics raise most often.

When a product gets more expensive but also better, statisticians try to strip out the portion of the price increase attributable to the improvement. This is called hedonic quality adjustment.

The logic is defensible. A car today costs more in nominal terms than one from 1995, but has far more safety equipment, uses less fuel and lasts longer. It isn't the same product.

The criticism is equally defensible. If I didn't want the extra quality and have to pay for it because the simpler model no longer exists, my purchasing power has genuinely fallen, whatever the adjustment says.

Both are true. It's a real methodological dilemma without a clean solution.

Shrinkflation

An effect that gets increasing attention: the package gets smaller and the price stays the same.

The index captures this in principle, because prices are computed per unit quantity. In perception, though, it doesn't land as a price rise at all but as a vague sense of having been cheated.

This is a case where the statistics are more accurate than perception — just in the opposite direction to usual.

What the figure is good for

There is one more distinction worth having, because it explains a lot of confused discussion. Inflation falling does not mean prices falling.

If the rate drops from eight percent to two, prices are still rising — just more slowly. The level reached during the high-inflation period stays where it is. For a household that felt the squeeze, the announcement that inflation has normalised does not correspond to any improvement they can observe, because their weekly shop has not become cheaper. It has merely stopped becoming more expensive quickly.

Actual falling prices, deflation, is rare and generally regarded by economists as a worse problem than moderate inflation, because it encourages postponing purchases and makes existing debt harder to service. So the thing most people intuitively want — prices going back to where they were — is not a policy goal anyone is pursuing. That gap between what the public hears and what is actually being targeted probably does more damage to trust in the figures than any methodological argument.

My conclusion after spending some time with this: the inflation rate is a decent measure of what it was built for — a control variable for monetary policy and an aggregate observation of price developments.

It is not a statement about your personal purchasing power, and it was never intended as one.

Anyone who wants to know how their own finances stand needs to look at their own spending structure. Some statistical agencies now offer personal inflation calculators where you enter your own expenditure. The result frequently diverges noticeably from the official figure, in both directions.

And that isn't a criticism of the statistics. It's just the difference between an average and an individual case — a difference we routinely gloss over in public debate.