Personal Money · Saturday, 15 August 2026

01 · Briefing · what happened

Inflation and the real return: why the number on your statement is not what you earned

Personal Money 5 min 18 sources

A savings account paying 4% while prices rise 3% earned you 1% in things you can actually buy. Here is how to read the only figure that answers the question.

3.4%

US consumer prices, year to July 2026

down from a 4.2% peak in May

0.40%

average US savings account rate

the biggest banks pay about 0.01%

18.65%

a three-month US savings certificate, Dec 1980

prices rose 12.5% that year, so the real gain was about 6 points

A$67,000

what A$100,000 buys today, in 2010 money

the number crossed six figures; the life it buys did not

At a glance

A savings rate means nothing on its own - the real return is your rate minus the rate prices rose.
US consumer prices rose 3.4% in the year to July 2026, while the average US savings account pays about 0.40%.
That average saver loses roughly 3% of purchasing power a year, and the balance never falls, so nothing records it.
A three-month US savings certificate paid 18.65% in December 1980, but prices rose 12.5% that year, so the real gain was about 6 points.
Pay follows the same sum: Australian wages rose 3.4% while prices rose 3.8%, so real wages went backwards.
Cash in a drawer earns 0%, so its real return is exactly minus the rate prices rose.
The official rate is an average of a basket built from choices - about 80,000 prices a month across more than 200 categories.
Different official measures of the same month disagree: 3.7%, 3.3%, 2.2%, 2.8% and 5.1% were all June 2026 US readings.

Forces in play

Rising prices High↓

US prices up 3.4% in the year to July 2026, easing from a 4.2% peak in May

What banks pay Easing→

the average US savings account pays about 0.40%, far below the rate prices are rising

The comfort of a steady balance Building↑

a balance that never falls feels safe, which is why the slow loss goes unnoticed

In play Savers — hold money at a rate that may sit below the rate prices rise Statistics agencies — build the basket and choose the methods that produce the official rate Central banks — target a rate of price rises - 2% in the United States

How it unfolded

Dec 1980 a three-month US savings certificate pays 18.65%, with prices rising 12.5%
Oct 2022 UK inflation peaks at 11.1% and no savings account beats it
May 2026 US prices up 4.2% on the year, the fastest in three years
Jul 2026 US prices up 3.4%, while the average account pays about 0.40%

Where this points

Watch the gap between what ordinary accounts pay and the rate prices rise - that single spread decides whether a steady balance is quietly gaining ground or losing it.

Full briefing

The question a rate cannot answer alone

Your bank quotes a rate. Say 4%. Is that good?

You cannot know yet. A rate is half a calculation. The other half is what prices did over the same year [1].

Work it through. Put $10,000 in an account paying 5%. A year later you have $10,500. But the car you were saving for cost $10,000 and now costs $10,300, because prices rose 3%. You buy the car and $200 is left over [1].

Your gain was not 5%. It was 2%.

That 2% is the real return: your rate minus the rate prices rose [1]. Nominal is the number printed on the statement. Real is what the money will actually buy. The quick rule is to subtract one from the other.

Today’s arithmetic

US consumer prices rose 3.4% in the year to July 2026 [3]. That eased from 3.5% in June [3][4]. In May the annual rate had hit 4.2%, the fastest in three years, as an energy shock fed through [5][3]. The Federal Reserve aims for 2% [4].

Now set that against what ordinary accounts pay. The national average US savings rate is about 0.40%, and the largest banks pay close to 0.01% [6]. The best-paying accounts sit around 4% to 5% [6].

So the average saver’s sum is 0.40% minus 3.4%. That is roughly minus 3% a year in purchasing power [6][3].

The balance never falls. What it buys does. Nothing on the statement records the loss.

The big number that was smaller than it looked

Run the same test backwards.

In December 1980 a three-month US savings certificate paid 18.65% [7]. Astonishing by today’s standards. But US consumer prices rose 12.5% over that year [4].

Subtract, and the real gain was about 6%, not 18.65%. Two-thirds of that eye-catching figure was simply prices moving.

Neither 18.65% nor 0.40% means anything until you know the number sitting beside it.

A pay rise can be a pay cut

The same arithmetic runs through wages.

Take the textbook case. Income goes from $50,000 to $52,000. That is a 4% rise. If prices rose 3% over the same year, the real gain is 1% [2].

Now the live version. Australian official figures showed wages up 3.4% over a year. Consumer prices over the same period rose 3.8%. Real wages went backwards [8]. Almost everyone got a rise. Almost everyone could buy slightly less.

Stretch it over a decade and the drift gets large. Adjusted for prices, A$100,000 today carries roughly the purchasing power of A$67,000 in 2010 [8]. About one in ten Australian full-time workers earned six figures in 2010. By 2025 it was around 45% [8].

The number crossed a threshold. The life it buys did not.

Economists measure real pay several ways, using different pay series and different price series, and the answer shifts with the choice [9]. Worth holding loosely.

Cash has no cushion

Money sitting in a drawer earns 0%. Its real return is exactly minus the rate prices rose.

Held over long stretches, cash rarely grows fast enough to hold its purchasing power [18]. That is not a claim about what anyone should do with an emergency fund. It is a description of the arithmetic.

The average is not your basket

Here is the honest caveat.

The official rate is not a measurement of your life. It is a weighted average of a basket, and the basket is built from choices.

The US Bureau of Labor Statistics samples about 80,000 prices a month across more than 200 categories of goods and services [10][11]. The data covers about 93% of the US population [11]. Housing is estimated from a survey of 50,000 rental units, including a modelled rent for owner-occupied homes that no owner actually pays [11][12].

Other decisions sit underneath. If a laptop costs the same as last year but runs twice as fast, the statistics count that as a price cut [12]. When steak gets dear and shoppers switch to chicken, the index adjusts for the swap [11][12]. Each choice is defensible on its own. Together they can produce a number at odds with what people feel at the till [12].

The measures also disagree with each other. For June 2026 the Fed’s preferred gauge read 3.7%. Strip out food and energy and the same report read 3.3%. A Dallas Fed measure that discards the biggest swings read 2.2%. An Atlanta Fed split put slow-moving prices at 2.8% and fast-moving ones at 5.1% [12]. Same country, same month, all defensible.

The plumbing shifts too. In early 2026 a quiet change of data source for legal services shaved roughly a tenth of a percentage point off a monthly core reading. Legal services are under 1% of consumer spending. There was no public notice; economists found out by asking [13].

So your own basket may be rising faster or slower than the headline. Rent in a hot city and the country’s number is not your number.

Why the illusion holds

Economists have a name for reacting to the figure rather than to what it buys: money illusion, after the American economist Irving Fisher [14]. One plain definition is the tendency to react to the number on the price tag or the payslip rather than to what the money will buy [15].

The economist Justin Wolfers puts the working test simply. Whenever you look at a dollar amount, whether a pay rise or an investment return, ask whether the inflation-adjusted figure would describe the trade-off better [14].

What varies, and what does not

Country and period change everything. In Britain, with inflation around 3.6%, roughly half of savings accounts beat it, according to the data firm Moneyfacts [16]. In October 2022, when UK inflation peaked at 11.1%, no savings account beat it at all [16].

Sometimes there is no rate on the shelf that holds purchasing power. That is a real state of the world, not a failure to shop around.

Some instruments are built around the problem. The US Treasury created inflation-protected securities in 1997, tied to an inflation gauge so the value moves with prices [17]. That is a mechanism, not a suggestion. How any of them work out depends on the price paid and the period held.

What does not vary is the arithmetic. Two numbers, never one.

The one thing to carry

A rate is half a sentence. Finish it with the rate prices rose, and you have the only figure that answers the real question: will this money buy more next year, or less?

02 · Lesson · why it matters

Why a number that never falls can still be shrinking

A balance that only ever grows can lose ground every year - what moved was not the number but what it buys.

How it works

Your account pays a rate - the nominal number
Prices rise by their own rate over the same year
Subtract one from the other for the real return
The balance still climbs, so nothing looks wrong
But the same money now buys a smaller basket

The twist

A number that never falls can shrink every year - because what moved was not the money but the price of everything it was for.

Where you've seen this

A pay rise

a 3% rise in a 4% year is a pay cut in everything but the payslip

A fixed pension

a payment that never falls buys less each year unless it is tied to prices

A long fixed-rate debt

the sum runs the other way - rising prices shrink what the borrower really owes

A salary you remember

the figure that once meant comfort measures a different life now

The catch

The official rate is an average of a basket built from methodological choices, so your own basket may be rising faster or slower than the headline number.

Full lesson

The statement that never goes down

Open a savings app and you see one figure. It was smaller last year. It will be larger next year. Nothing about it looks like a problem.

That is the trap, and it is built into the design. A bank statement records one thing: how many units of currency you hold. It has no column for what those units buy. The surface people check most often is the surface least able to tell them what happened.

A balance can rise every year and buy less every year. Those two facts sit together comfortably, because only one of them is ever displayed.

Two questions wearing one number

Every rate answers a question. The trouble is that people hear a different question than the one being answered.

The rate answers this: how many more currency units will you hold? The question people actually care about is this: will you be able to buy more?

Those are the same question only when prices hold still. They never do. So a rate on its own is an unfinished sentence. The missing half is not a footnote - it is often larger than the part you were shown.

That is why a huge rate in a high-inflation decade can be worth less than a modest rate in a calm one. The impressive figure was mostly compensation for something being taken at the same time.

The rise that was a cut

The same structure runs through pay, and it stings more there, because a pay rise arrives as a personal event.

Someone tells you a number went up. There was a conversation, maybe a negotiation. It feels like recognition. Meanwhile a separate, impersonal process has moved every price you face. The two are never discussed in the same room.

So the arithmetic can leave a person worse off in a year they were told they did well. The feeling that follows - running to stand still on a salary that used to signal comfort - reads as personal failure rather than as subtraction. The number went up. Something must be wrong with me.

Nothing is wrong with them. They were shown one half of a sum.

Somebody chose the basket

Now the part that is easy to miss, because it wears the clothes of a fact of nature.

“Inflation was 3.4%” sounds like a reading off an instrument, the way you would report a temperature. It is not. It is the output of a long chain of decisions somebody made.

Which goods go in the basket. How much weight each one carries. Whether a laptop that got faster at the same price counts as a price cut. What a homeowner with a fixed mortgage should be treated as paying in rent. Whether to follow shoppers when they switch from steak to chicken, or hold the basket fixed and record the steak.

Each of those has a defensible answer, and reasonable people chose them. But they are choices, and the number they produce is not neutral. It sets how benefits rise. It moves tax thresholds. It anchors what a pay negotiation treats as standing still. Change a method quietly and money has moved between people who never saw the decision.

That is the shape beneath the event. Not a conspiracy - a set of technical judgements that pose as measurement, which everyone downstream is then obliged to treat as the truth.

Nobody is holding their own number

Here is the humbling part, and it reaches the reader directly.

The official rate is an average, and an average is nobody. Someone who rents in an expensive city, drives a long commute and feeds small children lives in one basket. Someone with a paid-off house and a bus pass lives in another. Both are handed the same figure and told it describes their year.

So a person checking whether they are ahead is using a ruler built for a composite household that does not exist. Building their own is not really available. It would mean tracking every price they face, weighted by how much of it they buy. That is the job a statistical agency does with tens of thousands of price checks a month.

That is true of the saver and the wage-earner. It is also true of the officials setting rates off the same imperfect gauges, and of the economists still arguing about which gauge is least wrong.

Everyone in the chain is navigating by an average that fits nobody exactly, including the people who built it. The sum itself is simple. Knowing where you actually stand inside it is not.

03 · Lab · your turn

What It Actually Buys

Read a growing balance, commit to whether you are ahead, then see what the money will really buy.

04 · Hope · carry this

The arithmetic that quietly works against a saver is also the arithmetic anyone can learn in an afternoon. Once you know to ask for the second number, it stops being hidden.