The Inflation Lie: How CPI Understates the True Cost of Living

Author: GBTT Research

Category: Monetary Policy | Cost of Living

---

What Is the CPI — and What Does It Leave Out?

The Consumer Prices Index (CPI) is the UK government's official measure of consumer price inflation. It tracks the prices of a basket of goods and services, updated annually to reflect spending patterns. It is the measure the Bank of England uses to target 2% inflation.

CPI excludes several major costs that most working-age adults actually pay:

• Mortgage interest payments (MIPs) — excluded entirely from CPI
• Council tax — excluded from CPI
• Owner-occupiers' housing depreciation — excluded from CPI
• Buildings insurance — excluded from CPI

These are not trivial omissions. For most households, housing is the single largest expenditure category. The ONS itself acknowledges that owner occupiers' housing costs account for approximately 17% of the CPIH basket — but that 17% is simply absent from the headline CPI figure the government uses and reports.

The ONS publishes a broader measure, CPIH, which attempts to include owner-occupiers' housing costs using a method called "rental equivalence" — essentially estimating what a homeowner would pay in rent for their own property. This is an imputed, not a real, cost. It does not reflect rising mortgage payments when interest rates go up. It does not reflect the capital cost of buying a home. As of December 2025, CPIH ran at 3.6% against CPI's 3.4% — a modest gap that many economists argue still substantially understates the actual housing cost burden on households.

---

RPI: What It Includes That CPI Doesn't

The Retail Prices Index (RPI), first calculated in 1947, is the older and more comprehensive measure. It includes:

• Mortgage interest payments
• Council tax
• Owner-occupied housing depreciation
• House buildings insurance
• Ground rent

According to the OBR's working paper on the long-run RPI-CPI gap, "the RPI and RPIX include housing components such as owner-occupiers' housing depreciation and council tax and rates, which are currently excluded from the CPI. In addition the RPI further includes mortgage interest payments."

The result: RPI has run consistently higher than CPI. It ran above CPI in 22 out of 27 years between 1989 and 2015. In December 2025, RPI was 4.2% — 0.8 percentage points above CPI of 3.4%, and 0.6 points above CPIH of 3.6%.

The ONS officially discourages use of RPI. It states bluntly: "I believe that the RPI is not a good measure of inflation." That may be true in a narrow technical sense. But given who benefits from the lower CPI figure — see below — that discouragement deserves scrutiny.

---

The Formula Effect: Maths That Always Produces a Lower Number

Strip out every difference in coverage between CPI and RPI — remove housing, council tax, all of it — and CPI would still be lower. The reason is mathematical.

At the most granular level of price collection, CPI and RPI use different methods to aggregate individual price quotes into a sub-index:

• RPI uses the Carli formula (arithmetic mean of price relatives): adds price changes and divides by the count
• CPI uses the Jevons formula (geometric mean): multiplies price changes and takes the root

For any set of prices where there is variance between individual items, the geometric mean (Jevons) will always produce an equal or lower result than the arithmetic mean (Carli). This is a mathematical certainty, not a measurement judgment.

The OBR estimates the formula effect contributes approximately 0.9 percentage points to the long-run wedge between RPI and CPI. The Bank of England's own estimate of the total long-run wedge is 1.3 percentage points; the OBR's revised 2017 estimate is 1.0 percentage point; the UK Statistics Authority put it at approximately 1 percentage point.

The Carli method has been criticised for not meeting international standards. The Jevons method has been adopted internationally. Both observations may be simultaneously true: Carli can inject spurious upward bias in some circumstances, and Jevons produces a systematically lower number that benefits those who pay out inflation-linked obligations.

---

2003: Gordon Brown Switches the Target

Until December 2003, the Bank of England's inflation target was 2.5% measured by RPIX — the Retail Prices Index excluding mortgage interest payments. RPIX retained most of what RPI includes, except MIPs.

In his June 2003 pre-Budget report, Chancellor Gordon Brown announced the switch to a 2% CPI target. The rationale given was harmonisation with European norms — the UK used CPI (then called HICP) for cross-EU comparisons, so switching the domestic target would create consistency.

The effect was also entirely predictable: a lower-reading index, set against a lower numerical target (2% vs 2.5%), would allow monetary policy to remain looser for longer. As a Bank of England speech from October 2003 acknowledged, "the rate of inflation of HICP has on average run about three-quarters of a percentage point below that of RPIX." The new target was set lower precisely because the new index was lower.

Between May 1997 and December 2003, RPIX averaged 2.4%. CPI over the same period averaged materially less. The switch was not measurement neutrality. It was a deliberate choice of the lower measure.

---

2010: The Coalition Switches Pensions and Benefits

In July 2010, Pensions Minister Steve Webb announced that public sector pensions would be uprated annually in line with CPI rather than RPI. The change applied from April 2011. This was extended to the statutory minimum for private sector defined benefit pension increases and to working-age benefits.

The financial impact was immediate and substantial. A pensioner receiving £10,000/year, uprated at typical RPI (then projected at ~3.4%) rather than CPI (~2%), would accumulate approximately £284,923 over 20 years. Uprated at CPI, they would receive approximately £245,500 — a difference of £39,423 over a retirement.

The government's aggregate saving: the OBR estimated annual savings of £7.56 billion in 2014–15 rising to £10.6 billion in 2015–16. This figure covered public sector pensions, state benefits, and tax credits.

Private sector schemes could also switch. The Department for Work and Pensions calculated in July 2011 that private sector DB schemes would collectively save £3.342 billion per year from using CPI instead of RPI.

KPMG estimated the switch could reduce private sector pension liabilities by up to £100 billion.

The TUC described it as "a stealth cut on the pensions of middle income Britain." The CBI welcomed it. The government got £10 billion a year in savings. Pensioners and benefit recipients got lower annual increases, compounding year after year.

---

CPIH: The Partial Fix That Still Falls Short

In March 2017, the ONS made CPIH its "lead measure" of inflation, replacing CPI. CPIH includes owner-occupiers' housing costs — a meaningful improvement. But it uses rental equivalence: it estimates OOH costs by looking at what a homeowner would theoretically pay to rent a comparable property.

This methodology has a critical flaw. When house prices rise sharply but rents lag — which happens when property yields compress — rental equivalence underestimates the cost of housing. An owner who bought a house for £150,000 in 2010 and now has a house worth £300,000, with an LTV remortgage at current rates, faces mortgage payments that bear no resemblance to the local rental market equivalent.

The cost of actually buying a house — saving for a deposit, taking on a 25-year mortgage, absorbing higher interest rates — is simply not captured in any of the official indices. The CPIH OOH component rose 4.2% in December 2025. UK house prices rose over 200% between 2000 and 2024. These are not the same reality.

---

The Money Supply Problem

The most fundamental challenge to official inflation figures is the gap between money supply growth and reported price inflation.

In March 2009, the Bank of England launched quantitative easing in response to the Global Financial Crisis. By November 2020, it had purchased £895 billion of bonds — £875 billion in UK government gilts and £20 billion in corporate bonds — financed by newly created central bank reserves (digital money).

UK M4 broad money supply (notes and coin plus all sterling deposits with UK banks and building societies) now stands at approximately £3.2 trillion (December 2025), compared to approximately £1.4 trillion in 2009. That is a nominal increase of roughly 130% over 16 years.

Basic monetary theory — famously articulated by Milton Friedman — holds that "inflation is always and everywhere a monetary phenomenon." If you increase the money supply by 130% over 16 years, prices should broadly follow. The question is not whether they rose, but where.

The answer: not in the CPI basket. They rose in:

• UK house prices — up ~230% since 2000; up ~60% since 2009
• Equities — the FTSE 100 doubled between March 2009 and 2021; global equity markets tripled
• Bonds — gilt prices rose dramatically as yields were suppressed by QE itself
• Collectibles, art, prime property — all surged

CPI measures a basket of consumer goods and services: groceries, clothing, utilities, entertainment. It does not measure asset prices. The newly created money did not, for the most part, flow into the CPI basket. It flowed into assets. And assets are held disproportionately by the wealthy.

---

The Basket of Goods: What Goes In, What Comes Out

The ONS updates the CPI basket annually. Items are added and removed to reflect changing consumption patterns. This is methodologically defensible. It also creates systematic downward pressure on measured inflation through two mechanisms.

Substitution: When a product becomes expensive, consumers switch to cheaper alternatives. If the ONS then updates weights to reflect the substituted basket, inflation in the original basket is not fully captured. The index measures what people end up buying, not what they wanted to buy.

Hedonic adjustment: For products that improve in quality over time — electronics, vehicles — statistical agencies estimate how much of a price increase reflects quality improvement rather than inflation, and subtract it. ONS applies hedonic regression to laptops, PCs, tablets, smartwatches, and smartphones. In practice, the net effect of hedonic adjustments in the UK has been argued to be small (a 2006 US study found the net effect was approximately +0.005% per year), but the direction of the adjustment is always downward on measured inflation for technology goods.

Neither technique is straightforwardly dishonest. Both, however, consistently produce lower measured inflation than a fixed basket of what people actually bought last year. For households whose consumption patterns do not substitute — because they are on fixed incomes, or because the substituted product is inferior — measured CPI is not their inflation.

---

Who Benefits, Who Loses

The government benefits from lower CPI through:

• Lower gilt payments on index-linked bonds: The UK has approximately £600 billion of index-linked gilts outstanding. When payments are tied to CPIH instead of RPI — a switch announced for 2030 — the Treasury saves an estimated £2 billion per year.
• Lower benefit uprating: The switch from RPI to CPI for working-age benefits since 2011 has reduced the annual benefit bill by billions, with compounding effects year after year.
• Lower public sector pay settlements: When pay rises are benchmarked against CPI inflation rather than RPI or actual cost-of-living increases, real wages fall even when nominal wages rise.
• Lower inflation of tax thresholds: If allowances and thresholds rise with CPI, and real inflation is higher, the effective tax burden creeps up without the government having to legislate a rise.

Those who lose:

• Pensioners: Both state and occupational. Every year that actual price increases exceed CPI, their purchasing power erodes.
• Benefit recipients: Universal Credit, housing benefit, working tax credits — all uprated by CPI. If actual prices rise faster, the real value of benefits falls.
• Workers: Pay settlements anchored to CPI mean real wages decline when actual inflation runs higher.
• Savers: Cash savings grow at interest rates managed to a CPI target. If true inflation is higher, the purchasing power of savings falls.
• Anyone on a fixed income: The compounding effect is devastating. A 1 percentage point understatement of inflation sustained over 20 years produces a 20% cumulative shortfall in purchasing power.

---

The House Price Evidence

The most concrete evidence that CPI understates the true cost of living is UK house prices.

In Q1 2000, the average UK house price was approximately £80,000 (Nationwide data). By Q4 2024, it was approximately £265,000 — a nominal increase of 231% over 24 years.

Over the same period, cumulative CPI inflation was approximately 90%.

A person who needed to buy a house in 2000 but couldn't, and is trying to buy the same house in 2024, has experienced ~231% inflation in that specific cost. CPI says their purchasing power has only eroded by ~90%. The gap is not theoretical. It is the difference between being able to buy a house and not.

The counterargument is that house purchases are investments, not consumption expenditure. But people do not buy houses as investments; they buy them because they need somewhere to live. The cost of accessing shelter has risen three times faster than CPI. That is not captured in any measure the government uses to uprate wages, benefits, or pensions.

---

Reading the Numbers

As of December 2025:

Measure12-Month RateWhat It Includes CPI3.4%Consumer goods and services; no housing costs CPIH3.6%CPI + imputed rental equivalent for owner-occupiers + council tax RPI4.2%CPIH items + mortgage interest payments + housing depreciation

The gap between CPI (3.4%) and RPI (4.2%) is currently 0.8 percentage points. The OBR's long-run assumption for the wedge is 1.0 percentage point. The formula effect alone — pure maths — accounts for 0.9 percentage points.

A worker whose pay rises were set against CPI for the past five years has seen their nominal pay increase track 3–4% inflation. But their mortgage payments, council tax, and actual cost of occupying a house have tracked something closer to RPI — or, in the case of house prices themselves, far beyond any inflation measure.

---

Conclusion

The inflation measure the government targets, reports, and uses to set benefits, pensions, and pay is one that excludes the two largest costs for most households — housing and council tax. It uses a mathematical formula that produces a lower number than the alternative by design. It was switched to from a higher-reading measure in 2003 because it was lower. It was used to replace pension and benefit uprating in 2010 because it was lower. And the Treasury saves billions each year precisely because it is lower.

None of this proves a conspiracy. Each individual methodological choice has a technical rationale that statisticians can defend. But the aggregate effect is a measure that has consistently produced lower readings than people's actual cost of living, and that has been repeatedly chosen, over alternatives, by the same institution that benefits financially from the lower number.

The measure is not wrong in a crudely fraudulent sense. It is designed to answer a specific narrow question — how much did a defined basket of consumer goods change in price? — and it answers that question reasonably well. The problem is it is used to answer a much broader question — how much has your cost of living risen? — for which it is structurally unsuited.

When the Bank of England created £895 billion and house prices rose 60% in a decade and CPI barely moved, the instrument was not measuring the thing people care about. That gap between official inflation and lived experience is not a rounding error. It is a policy choice with winners and losers — and the government is firmly in the winner's column.

---