The claim is usually true and it is not true by definition, and the reasons it holds are worth separating from the reasons it might not.
The case that the poor are harmed more begins with what they buy. Poorer households spend a much larger share of their income on food, fuel, housing and transport, and these necessities frequently rise faster than the general price level — so the inflation a poor household actually experiences can be several points above the published figure, which is an average over a typical basket. They also cannot substitute away from a price rise: a wealthy household facing dearer fuel can drive less or insulate its home, while a household already at the minimum has nothing to cut.
The second reason is the form their wealth takes. Poorer households hold what savings they have as cash or bank deposits, which lose real value pound for pound as prices rise. Wealthier households hold property, shares and businesses — real assets, whose prices tend to rise with inflation, so their wealth is largely protected and may even increase. Inflation therefore erodes the savings of those with least and leaves the assets of those with most intact.
The third is bargaining power. Workers in strong unions and in scarce occupations negotiate pay rises that keep up with prices; the low-paid, the non-unionised and those on benefits uprated once a year and in arrears do not. So the incomes best protected against inflation belong to those who needed protection least.
The case that it is not always so rests on debt. Inflation reduces the real value of fixed debts, and borrowers gain accordingly — a household with a large fixed-rate mortgage can gain substantially. Wealthy households and firms are often the largest borrowers in absolute terms, but a mortgage is also the main asset-linked debt of middle-income households, so the transfer does not run purely by income. Those who benefit least from this channel are those too poor to borrow at all. It is also worth noting that where benefits and pensions are index-linked, the poorest may be protected more systematically than middle-income workers whose wages are not.
My judgement is that inflation harms poorer households more, and that the decisive reason is the composition of what they own rather than what they buy. The spending argument is real but partly correctable: a government can target help at food and fuel costs, and can index benefits more frequently. The asset argument is structural. As long as the poor hold cash and the rich hold property and shares, inflation transfers real wealth upwards automatically, and no adjustment to benefits alters that.
The qualification about debt is genuine and limited. It benefits those who have been able to borrow, which in most countries excludes the poorest entirely — so the exception applies mainly to the middle rather than to the bottom.
This has a clear policy implication: if inflation is allowed to persist, index-linking benefits and pensions addresses only part of the harm. Protecting the savings of those without assets requires accessible index-linked saving, and preventing the wealth transfer altogether requires preventing the inflation.