Has Teacher Pay Fallen Since 2010? Why the Numbers Disagree

The four numbers, side by side

Almost every argument about teacher pay since 2010 is really an argument between people quoting different statistics at each other without saying which one they are using. Here they are together.

Look at the last two columns. Every difference in the second column is explained by a difference in the third, fourth or fifth. Nobody in that table is making anything up.

The three things that make the numbers differ

There are exactly three methodological choices behind every published figure, and once you know which choice each source made, you can translate between them.

• Which inflation index deflates the cash figures. CPI, CPIH and RPI produce materially different answers over fifteen years, and the gap compounds.
• What is being measured. A fixed point on the pay scale is a different quantity from the median earnings of the actual workforce, because the workforce itself changes.
• Which years are compared. 2010/11 to 2024/25 is a different span from 2010/11 to 2025/26, and a figure that stops before the most recent award will always look worse than one that includes it.

Axis 1: CPI, CPIH or RPI — and why it matters so much

The most consequential choice is the inflation index, and it is the one least often stated in headlines.

The three indices measure overlapping but different baskets and, more importantly, use different mathematics to average price changes. The practical consequence is a persistent wedge. Taking the STRB's own reported figures for the twelve months to January 2026: CPI rose 3.0%, CPIH rose 3.2%, and RPI rose 3.8%. That is 0.8 percentage points between the lowest and the highest, in a single ordinary year.

A large part of that wedge is the formula effect — the difference in the annual rate of change caused purely by how price averages are calculated, principally RPI's use of the Carli formula. The gap attributable to it was around 0.5 percentage points before 2010 and has been around 0.8 percentage points since. It is not a reflection of teachers facing different prices; it is a property of the arithmetic.

Is RPI therefore the wrong index? The honest answer is more nuanced than either side usually allows.

Axis 2: fixed scale points versus median actual earnings

The second choice is subtler and, for anyone trying to understand their own pay, arguably more important than the inflation index.

A fixed scale point comparison takes a specific rung of the ladder — M1, M6, the leadership minimum — and asks what the cash rate for that rung has done relative to prices. It holds the job constant. It is the cleanest possible measure of what pay award decisions did, because nothing else in the comparison changes.

A median actual earnings comparison asks what the typical teacher actually earned. That number moves for two quite different reasons: because pay rates changed, and because the mix of people changed. If the workforce gets younger and less experienced, median earnings fall even if not a single pay point moved.

This is not hypothetical. The STRB's own report records that the proportion of teachers aged 50 or over fell from 24% in 2010/11 to 20% in 2023/24, and that the proportion of leavers who were retiring fell over the same span. A workforce weighted more towards early-career teachers has a lower median by construction.

That composition effect is very likely why the STRB's median figure (−13.9%) is larger than any of its own fixed-point figures (−1.0% to −11.2%). The two are not in conflict; they are measuring different things, and the difference between them is itself informative — it is roughly the size of the workforce-composition effect.

Axis 3: start and end dates

The third choice is the least interesting intellectually and the most abused rhetorically.

2010 is the shared reference point in this debate because it marks the start of the post-financial-crisis public sector pay restraint period — a pay freeze followed by a multi-year pay cap. Both government and union evidence to the STRB use it, and they draw opposite conclusions from it, which is a good sign that the baseline itself is not the thing in dispute.

The end date matters more than people assume, because the last few years have been unusually volatile. A series ending in 2022/23 captures the peak inflation year and looks terrible; one ending in 2025/26 includes several above-inflation awards and looks considerably better. The STRB's scale point table ends at 2025/26. Its median earnings figure ends at 2024/25, because the School Workforce Census data it relies on lags. Neither includes the 3.5% award from 1 September 2026.

And a figure can use a completely different baseline altogether. The Department for Education's line that teachers "will see an increase in their pay of 17% since this government took power — equating to almost £7,900 over four years" is a cumulative cash figure over a four-year window beginning in 2024. It is not a real-terms comparison and it is not measured from 2010. It is not dishonest; it is answering a different question, and it is perfectly possible for it and the 13.9% real-terms fall to be simultaneously true.

The STRB's scale point table, in full

Because this is the most authoritative single source — produced by the body that recommends teacher pay, using the department's own data, published July 2026 with data up to 18 February 2026 — it is worth reproducing exactly as printed. Table B.7 of the 36th Report gives real-terms change from 2010/11 to 2025/26, based on the STPCD, with nominal earnings adjusted using annual financial-year CPI figures:

Two notes on reading it honestly. First, the Upper Pay Range row is printed in the report as "UPR1/UP3", a label the report does not expand; we have reproduced it as published rather than guessing which points it covers. Second, the STRB's own narrative summary of the table is that "whilst all pay points and ranges have declined in real terms, M1 declined the least (by 1.0%) whilst the leadership maxima declined the most (by 11.2%)".

Why M1 held up and the top of the profession did not

The 1.0% figure for M1 surprises people who expect a uniform story of decline, and it is worth understanding because it is the result of deliberate policy rather than accident.

Successive pay awards have been weighted rather than flat. In several recent rounds, government and the STRB directed larger percentage increases to the bottom of the Main Pay Range, for two connected reasons: acute early-career recruitment shortfalls, particularly in shortage subjects, and the rising National Living Wage, which creates an obvious problem if a graduate profession's starting salary drifts towards it. The 2026/27 award continues the pattern — 3.5% across all pay and allowance ranges, but 5% specifically on the minimum of the unqualified teacher range for the rest of England, which the STRB explicitly framed as mitigating the risk of falling below the National Living Wage.

The cumulative effect is pay compression: the real-terms gap between starting pay and experienced pay is narrower than it was in 2010. That is a defensible trade if the binding constraint is getting people into the profession. It is considerably less comfortable if the binding constraint is keeping them, because the financial reward for staying, gaining experience and taking on responsibility has shrunk relative to where it was. The STRB's benchmarking against comparable market roles reflects the same shape: starting pay sits between the lowest quartile and the median for roles of comparable size and scope, becomes less competitive for teachers further up the Main Pay Range, and improves again on the Upper Pay Range.

If you are weighing whether to apply for the Upper Pay Range or take on a responsibility allowance, this is the real-world context: the structure rewards progression less in real terms than it did fifteen years ago. Our guides to the Upper Pay Range threshold and to TLR payments deal with the individual decision.

The comparison that matters most, and is quoted least

Real-terms pay against prices tells you about purchasing power. It does not tell you whether teaching has become a less attractive choice relative to the alternatives — and for recruitment and retention, that is the question that actually bites.

The STRB addresses it directly. Its analysis finds that teachers' median gross earnings in 2024/25 were 13.9% below their 2010/11 level in real terms. Over the same period, median gross earnings for the whole economy in England were 0.6% below their 2010/11 level, and median gross earnings for professional occupations in England were 9.2% below.

The STRB's own conclusion is that the competitiveness of teachers' earnings compared with the whole economy, the wider public sector and professional occupations was lower in 2024/25 than in 2010/11. Its report also charts a public sector line, though it does not quote a headline percentage for it in the same passage, so we have left that row out rather than estimate it from a chart.

NFER's analysis frames the same point in a way many teachers find more concrete. It finds that growth in teachers' starting salaries in real terms has been around seven percentage points lower than average earnings growth over 2010/11 to 2025/26, and growth in experienced teachers' salaries around sixteen percentage points lower. Expressed in cash, NFER estimates that had teacher pay kept pace with average earnings growth, the starting salary would have been about £2,500 higher in 2025/26 and experienced teacher pay about £9,200 higher.

NFER, the IFS and the union figures

NFER

NFER published its analysis on 15 December 2025, ahead of the 2026 pay round. It deflates using a combination of CPIH and the OBR's latest CPI forecast. Its headline findings are that a teacher's starting salary in 2025/26 was similar in real terms to 2010/11, while experienced teacher pay remained around nine per cent lower. That sits comfortably between the STRB's M1 figure of −1.0% and its M6 figure of −6.7%, which is what you would expect from a slightly different deflator and slightly different representative points.

The IFS

The Institute for Fiscal Studies has produced a long-running series on teacher pay using CPIH, generally finding steeper falls for more experienced teachers than for new entrants — reported figures in the region of a 5% fall for starting salaries and around 12 to 13% for more experienced and senior teachers at the low point in 2022/23, with a projection that pay for most teachers would still be around 7% lower in real terms in 2027/28 than in 2010/11. We have taken these from reporting of the IFS analysis rather than from the IFS pages themselves, which we could not retrieve directly, so treat the precise figures as indicative and check the IFS source before quoting them.

NEU and NASUWT

The NEU's figure is around 23%, measured against RPI, and it also expresses the loss cumulatively in cash: a teacher who started in 2010 and progressed up the scale is said to have lost close to £109,000 in real terms over their career to date. NASUWT publishes comparable RPI-based figures, including a figure of around 29.9% for most classroom teachers in Northern Ireland in 2025/26 — a reminder that the four nations have diverged and that a figure for one is not a figure for another. As above, these rest on RPI; the substantive disagreement with the STRB is about the deflator, not about the underlying cash series.

Which number to use for what purpose

Rather than force a single answer, here is the honest mapping between question and appropriate statistic.

The one thing worth refusing to do is to quote a percentage without stating the index, the measure and the years. A number without those three things is not a fact, it is a position.

Sources, and what we could not verify

• School Teachers' Review Body, Thirty-Sixth Report 2026 (PDF) — Table B.7 and Appendix B paragraphs 38 and 39 (scale point and median earnings real-terms figures), Appendix B paragraphs 3 and 4 (CPI, CPIH and RPI for the twelve months to January 2026), paragraphs 2.15 to 2.17 (inflation forecasts and relative pay), and the workforce composition data on teachers aged 50 or over. All figures attributed to the STRB on this page were read directly from this document.
• School Teachers' Review Body 36th Report: 2026 — GOV.UK (report and Government response).
• What should happen to teachers' pay from 2026? — NFER, 15 December 2025.
• UK Statistics Authority statement on the future of the RPI — ONS, and Calculating the Retail Prices Index — ONS (formula effect, loss of National Statistic status in March 2013, alignment with CPIH from February 2030).
• Teacher pay: everything you need to know about the latest multi-year pay award — DfE Education Hub (the 17% and £7,900 figures).
• Teachers' pay — NEU Edufacts and How much pay have teachers in Northern Ireland lost? — NASUWT.
• What has happened to teacher pay in England? — Institute for Fiscal Studies.

What we could not verify directly. The NEU, NASUWT and IFS pages listed above are behind bot protection and could not be retrieved for direct quotation; the figures attributed to them here come from search results and secondary reporting of their published analysis, and are labelled in the text as attributed rather than verified. If you intend to quote them in a formal setting — a pay case, a governors' paper, a union submission — go to the source pages yourself first. The STRB, NFER, ONS and DfE figures on this page were read directly from the documents linked above.

The reconciliation we cannot complete. It is not possible to reconcile the STRB and union figures to the pound, because the underlying cash series, the exact pay points chosen and the precise deflator vintages are not all published in comparable form. The 12.7 percentage point compounding illustration above shows that the inflation index alone plausibly accounts for most of the gap; it is an explanation of the direction and rough magnitude, not an exact bridge between the two numbers.