Decoded, Economics & finance

Taxing the Shock Absorber: Who Really Pays Britain’s Bank Surcharge?

Decoded analysis cover: Taxing the Shock Absorber
The British Economy in 2026 | Analytical Series, Paper 4

Read offline: the full paper is available as a PDF (9 pages).
Download PDF

Executive Summary

Who pays a tax on bank profits? Banks argue that higher taxes are passed on to borrowers and savers and weaken lending. Supporters of the tax argue that it falls on shareholders whose profits have been inflated by high interest rates. Both sides cite the same evidence: interest rates. This paper asks what the evidence can actually show.

  1. The 2027 increase is small relative to the banking system. Each percentage point of surcharge raises about 400 million pounds a year. Even if every penny were added to the price of mortgages, the 3-point increase would raise mortgage rates by about 8 basis points, or about 115 pounds a year on a 150,000 pound mortgage. On a wider base of loans and deposits the effect would be smaller still.
  2. Seventeen years of Bank of England data show no detectable link between the surcharge and lending margins. Regressions of nine lending and deposit margins on the surcharge rate give coefficients for mortgages between minus 3 and plus 0.4 basis points per percentage point of tax, with standard errors of 2 to 4 basis points.
  3. Event studies around the two previous changes find movements in the expected direction for standard variable rate mortgages, plus 7 basis points when the surcharge was introduced in 2016 and minus 13 when it was cut in 2023, but placebo tests at 122 other dates show that such jumps are entirely ordinary.
  4. The reason is statistical power. The smallest jump in margins that could be detected with 80 per cent power is 32 basis points for standard variable rates, 50 for instant-access savings and about 80 for fixed-rate mortgages. The largest plausible effect of the surcharge is 21 basis points for the 2016 introduction and 8 for the 2027 increase. Interest rate data cannot answer the incidence question.
  5. The one market signal available, a 3 to 4 per cent rise in bank shares when the Budget increase turned out smaller than the feared return to 8 per cent, suggests that investors expect shareholders, not customers, to bear most of the tax.

The policy implication is that the debate about the surcharge should move away from interest rates, where it cannot be settled, toward profits, dividends and lending volumes, which the surcharge may affect more visibly over time, and toward the question of whether a temporary tax on a cyclical windfall is a sound way to fund energy support.

1. Banks as the Payer of Last Resort

On 28 October 2026 the Chancellor announced that the bank corporation tax surcharge would rise from 3 to 6 per cent from 1 January 2027 for a period of fifteen months, returning to 3 per cent on 1 April 2028. The measure is intended to help fund a household energy payment of about 2.6 billion pounds. With the main rate of corporation tax at 25 per cent, banks with taxable profits above the 100 million pound allowance will pay a combined rate of 31 per cent during the temporary period.

Figure 1. The tax rate on bank profits, 2014 to 2029
Figure 1. The tax rate on bank profits, 2014 to 2029. Source: HM Treasury and HMRC. Main rate of corporation tax plus the bank surcharge on profits above the allowance (25 million pounds until March 2023, 100 million pounds since).

The surcharge has a short but eventful history. It was announced in July 2015 and introduced in January 2016 at 8 per cent, as the bank levy on balance sheets was being reduced. In April 2023, when the main rate of corporation tax rose from 19 to 25 per cent, the surcharge was cut to 3 per cent and the allowance raised from 25 to 100 million pounds, leaving the combined rate for large banks almost unchanged at 28 per cent. The 2027 increase is the first since 2016 to raise the overall rate.

The increase was smaller than many had feared. In the weeks before the Budget, reports that the surcharge might return to 8 per cent had weighed on bank shares. When the 6 per cent rate was confirmed, shares in the main British banks rose by 3 to 4 per cent. UK Finance, the industry body, had warned that a return to 8 per cent could cost 6 billion pounds of output over five years.

Table 1. The surcharge in numbers

Item Value
Surcharge rate, January 2016 to March 2023 8 per cent above a 25 million pound allowance
Surcharge rate, April 2023 to December 2026 3 per cent above a 100 million pound allowance
Surcharge rate, January 2027 to March 2028 6 per cent
Revenue raised at 3 per cent, 2025-26 about 1.2 billion pounds
Revenue per percentage point about 400 million pounds a year
Expected yield of the temporary increase about 1.5 billion pounds in total
Household energy payment it helps fund about 2.6 billion pounds

Sources: HM Treasury, Budget 2026 policy paper on the banking surcharge; press reports of the Budget costings.

2. How Much Could Pass Through?

Before testing whether the surcharge is passed on, it is useful to ask how large the effect could be. A tax on profits does not, in a textbook competitive market, change the price that maximises profit; it reduces the profit. Pass-through to customers can occur only if banks have market power and use the tax as a coordination point to raise prices, or if the tax raises their cost of equity and they price loans to maintain a target return.

An upper bound is obtained by assuming that the entire revenue is recovered from a single product. Each percentage point of surcharge raises about 400 million pounds a year. Spread over the 1,563 billion pounds of outstanding mortgages, that is 2.6 basis points of interest per percentage point of tax. The 2027 increase of 3 points would therefore raise mortgage rates by about 8 basis points if borne entirely by mortgage borrowers, and the 2016 introduction at 8 per cent by about 21 basis points. If the burden were spread across all household and business loans and across deposit margins, the effect on any single rate would be a fraction of this.

Table 2. The maximum effect of full pass-through to mortgage borrowers

Change Rate change (points) Effect on mortgage rates (bp) Annual cost on a 150,000 pound mortgage
Introduction, January 2016 +8 about +21 about 310 pounds
Cut, April 2023 -5 about -13 about -195 pounds
Temporary increase, January 2027 +3 about +8 about 115 pounds

Note: 400 million pounds per percentage point divided by the September 2026 mortgage stock of 1,563 billion pounds. Upper bound: assumes the entire tax is recovered from mortgage borrowers.

3. What the Data Show

3.1 Margins over seventeen years

The analysis uses monthly Bank of England quoted interest rates from January 2010 to October 2026 for nine products: two-year and five-year fixed-rate mortgages at 75 per cent loan-to-value, two-year fixed at 60 and 90 per cent, the standard variable rate, personal loans of 10,000 pounds, credit cards, instant-access savings and one-year fixed-rate bonds. Each is converted into a margin over the relevant risk-free rate: the zero-coupon gilt yield of matching maturity for fixed-rate products, Bank Rate for variable-rate products. For deposits the margin is the risk-free rate minus the deposit rate, so that a higher margin means savers receive less.

Figure 2. Lending and deposit margins, 2010 to October 2026
Figure 2. Lending and deposit margins, 2010 to October 2026. Source: Bank of England quoted household interest rates (IUMBV34, IUMBV42, IUMTLMV, IUMB6VJ) and zero-coupon gilt yields; author’s calculations. Shaded: surcharge at 8 per cent.

Figure 2 shows four of these margins. Their movements are dominated by monetary policy and market stress, not by tax. Fixed-rate mortgage margins collapsed in 2022 as swap rates spiked after the September mini-budget, then rebounded. Deposit margins were close to zero while Bank Rate was near its floor and widened sharply as rates rose. The standard variable rate margin, by contrast, has been remarkably stable.

Table 3. Margins and the surcharge rate: regression estimates, January 2010 to October 2026

Margin Coefficient on surcharge (bp per point) Standard error 95% upper bound
2-year fixed mortgage, 75% LTV -2.8 3.7 4.5
5-year fixed mortgage, 75% LTV -0.1 2.9 5.6
2-year fixed mortgage, 60% LTV 0.4 2.4 5.1
Standard variable rate 0.4 1.6 3.6
Personal loan, 10,000 pounds -7.2 10.1 12.6
Credit card -14.0 5.0 -4.2
Instant-access savings 6.5 2.8 12.0
One-year fixed bond 5.9 2.8 11.4

Note: each margin regressed on the surcharge rate, Bank Rate, the slope between the two-year gilt yield and Bank Rate, a lower-bound dummy, pandemic and 2022 gilt-crisis dummies and a linear trend. HAC standard errors (12 lags). For deposits, a positive coefficient means savers receive less.

Table 3 reports regressions of each margin on the surcharge rate, controlling for the level of Bank Rate, the slope of the yield curve, the effective lower bound, the pandemic, the 2022 gilt crisis and a time trend. For mortgages the coefficients are close to zero, and the upper ends of the confidence intervals, 4 to 6 basis points per point of tax, are only about twice the full pass-through benchmark of 2.6. For deposit margins the coefficients are positive and statistically significant, around 6 basis points per point, which would suggest that savers bore part of the tax; but deposit margins are dominated by the level of interest rates, and the 8 per cent era coincided almost exactly with the years of near-zero rates, so this association cannot be given a causal reading. The negative coefficient for credit cards has the wrong sign and is equally unlikely to be causal.

3.2 The two natural experiments

A sharper test looks at what happened to margins in the twelve months either side of each change, controlling for Bank Rate and the slope of the curve. If banks passed the tax on, margins should have risen when the surcharge was introduced in January 2016 and fallen when it was cut in April 2023.

Figure 3. Margins around the introduction of the surcharge, January 2016
Figure 3. Margins around the introduction of the surcharge, January 2016. Source: Bank of England; author’s calculations. Change in each margin relative to December 2015.

Table 4. Jumps in margins at the surcharge changes, with placebo p-values

Margin Jan 2016 (bp) p-value Apr 2023 (bp) p-value
Standard variable rate +6.9 0.50 -12.6 0.22
Instant-access savings +8.4 0.46 -38.2 0.11
One-year fixed bond +33.4 0.19 +49.0 0.11
Credit card +7.9 0.78 +52.1 0.21
2-year fixed mortgage -3.7 0.91 -110.5 0.00
5-year fixed mortgage -21.9 0.43 -71.7 0.07
Personal loan -15.9 0.48 -24.8 0.39

Note: jump estimated by regression on a post-change dummy, Bank Rate and the yield-curve slope over a 24-month window. The p-value is the share of 114 to 122 placebo dates, at least twelve months from any surcharge event, with a jump at least as large in absolute value. The April 2023 fixed-rate mortgage jumps reflect the unwinding of the 2022 gilt crisis.

The results in Table 4 look, at first sight, like evidence of pass-through for the least competitive products. The standard variable rate margin rose by 7 basis points when the surcharge was introduced and fell by 13 basis points when it was cut, exactly the pattern expected if banks recover the tax from inert customers who do not shop around. The instant-access savings margin also rose in 2016. Raw margins on fixed-rate mortgages also widened in early 2016, as Figure 3 shows, but only because gilt yields fell faster than mortgage rates; once the slope of the yield curve is controlled for, the jump disappears.

Figure 4. Were the surcharge changes unusual? A placebo test
Figure 4. Were the surcharge changes unusual? A placebo test. Source: author’s calculations. Distribution of estimated jumps in the standard variable rate margin at 122 dates at least twelve months away from any surcharge event.

The placebo test in Figure 4 shows why that reading cannot be sustained. Estimating the same jump at 122 other dates when nothing happened to the surcharge produces jumps of similar or larger size half the time for the 2016 introduction and a fifth of the time for the 2023 cut. None of the event jumps in Table 4 is statistically unusual, except the collapse of fixed-rate mortgage margins in April 2023, which reflects the normalisation of swap markets after the 2022 gilt crisis, not the surcharge.

4. Why the Question Cannot Be Settled With Prices

Figure 5. Why the surcharge is invisible in interest rates
Figure 5. Why the surcharge is invisible in interest rates. Source: author’s calculations. Minimum detectable jump: 2.8 times the standard deviation of placebo jumps (5 per cent two-sided test, 80 per cent power).

The underlying problem is one of statistical power. The standard deviation of placebo jumps measures how much margins move in an ordinary year. A change in margin can be detected with 80 per cent probability only if it is about 2.8 times that standard deviation. For the standard variable rate this minimum detectable jump is 32 basis points; for instant-access savings 50; for fixed-rate mortgages around 80; for credit cards over 90.

Every one of these thresholds is above the largest plausible effect of the surcharge. Full pass-through of the 2016 surcharge to mortgage borrowers alone would have raised mortgage rates by about 21 basis points; full pass-through of the 2027 increase would raise them by 8. Interest rate data are simply too noisy to reveal effects of this size. A finding of no effect in price data is therefore not evidence that banks absorb the tax, and the absence of any visible increase after 2016 is not evidence that customers were spared. Neither side of the debate can rely on interest rates.

Table 5. Minimum detectable effects and plausible effects (basis points)

Margin Minimum detectable jump Full pass-through, 2016 Full pass-through, 2027
Standard variable rate 32 21 8
Instant-access savings 50 21 8
2-year fixed mortgage 79 21 8
5-year fixed mortgage 81 21 8
Credit card 94 21 8

Note: full pass-through assumes the whole tax is recovered from a single product with a base equal to the mortgage stock; it is an upper bound for every product.

5. What Other Evidence Says

If prices cannot reveal the incidence of the surcharge, other evidence must be used. Three kinds are available.

The first is the stock market. When the Budget confirmed a 6 per cent surcharge instead of the feared 8 per cent, shares in the main British banks rose by 3 to 4 per cent. Investors priced the tax as a cost to shareholders: if they had expected banks to pass it fully to customers, a smaller-than-feared increase would have made little difference to share prices. Market prices before and after the announcement therefore point to shareholders as the main payers, at least in the short run.

The second is the structure of the tax. The surcharge applies only to profits above the allowance, now 100 million pounds per group, and only to banks and building societies. Smaller lenders below the allowance, and non-bank lenders, are not affected. In markets where they compete actively, such as fixed-rate mortgages, any attempt by large banks to raise prices would lose them market share. Pass-through is most plausible in products where customers are inert and competition is weak, such as standard variable rates and instant-access savings.

The third is the temporary nature of the 2027 increase. A tax that lasts fifteen months and is legislated to end is unlikely to change the cost of equity that banks use to price long-term products. Banks have little reason to reprice five-year mortgages for a tax that will have expired before the end of the first year of the loan.

Table 6. Incidence scenarios for the 2027 increase

Scenario Who pays Cost per year Illustration
Shareholders bear all Bank shareholders about 1.2 billion pounds Lower post-tax profits and dividends
Full pass-through to mortgages Mortgage borrowers about 8 bp about 115 pounds a year on a 150,000 pound loan
Pass-through to inert customers SVR borrowers and instant-access savers a few bp on those products Unlikely to be detectable

Note: the evidence in this paper cannot discriminate between these scenarios using interest rates; market reactions favour the first.

6. Lessons for Tax Policy

Three lessons follow.

First, the public debate about the surcharge relies on a kind of evidence that cannot answer the question. Claims that the surcharge has raised mortgage rates, and claims that it has not, are equally unsupported by interest rate data, because the effect is too small to be seen. Policy makers should not accept either claim on the basis of price charts.

Second, the better test of the surcharge is in quantities and profits: lending volumes to households and small firms, the market share of lenders below the allowance, dividends and buybacks. These are where the burden of a profits tax ultimately shows up, and they should be monitored during the fifteen months of the higher rate.

Third, the use of a temporary bank tax to fund energy support links two cyclical windfalls. High interest rates raised bank profits; high gas prices raised household bills. Taxing the first to offset the second is defensible as a one-off, but the third paper in this series shows that the energy shock will peak in household bills in the spring of 2027, so timing the payment to coincide with the January and April price-cap increases matters more than the exact source of revenue.

7. Conclusions

The doubling of the bank surcharge in 2027 will raise about 1.5 billion pounds over fifteen months. Whether it is paid by shareholders, borrowers or savers cannot be determined from interest rate data: even full pass-through would move rates by less than a tenth of the ordinary noise in bank margins. The two previous changes to the surcharge, in 2016 and 2023, left no statistically detectable trace in seventeen years of Bank of England data. The one market signal available suggests that investors expect shareholders to bear the tax. This is a negative result, but it should change the terms of the debate: the incidence of bank taxes has to be studied through profits and quantities, not through prices.

Methodological Appendix

A1. Regressions

For each margin m, m(t) = a + b S(t) + c1 B(t) + c2 [y2(t) – B(t)] + c3 ZLB(t) + c4 COVID(t) + c5 LDI(t) + d t + e(t), where S is the surcharge rate, B Bank Rate, y2 the two-year zero-coupon gilt yield, ZLB a dummy for Bank Rate below 0.6 per cent, COVID a dummy for March 2020 to June 2021 and LDI for September to December 2022. Monthly data, January 2010 to October 2026, Newey-West standard errors with 12 lags.

A2. Event studies and placebo tests

For each event date E, the margin is regressed on a dummy equal to one from E onwards, Bank Rate and the curve slope, over the 24 months from E-12 to E+11. The same regression is estimated for every month from January 2011 to October 2025 at least twelve months away from July 2015, January 2016 and April 2023, giving 114 to 122 placebo jumps per margin. The placebo p-value is the share of placebo jumps at least as large in absolute value as the event jump. The minimum detectable effect is 2.8 times the standard deviation of the placebo jumps.

A3. Limitations

Quoted rates are averages across lenders and do not distinguish lenders above and below the allowance; lender-level data would allow a difference-in-differences design with small lenders as a control group, and would raise power considerably. The upper bound for pass-through uses the 2026 mortgage stock and current revenue estimates for all periods. The stock market evidence relies on press reports of the share price reaction and is not a formal event study.

References

Bank of England (2026), Bankstats: quoted household interest rates, accessed November 2026.

Bank of England (2026), Yield curve data: nominal zero-coupon gilt yields.

HM Treasury (2026), Budget 2026; Increase in rate of banking surcharge for temporary period, policy paper, 28 October 2026.

UK Finance (2026), Pre-Budget representation on bank taxation.

About this series. This paper belongs to a ten-part series on the British economy in 2026 published by Global Strategy Files. Each study takes one current economic question and answers it with original time-series analysis and modelling on the latest official data.

Driss EFFINA, PhD
Independent Researcher, Global Strategy Files

Free preview
The Q4 2026 Global Outlook

The opening pages of our four country reports: United States, United Kingdom, Japan and China. A free PDF, sent to your inbox.

You will also receive occasional analysis from Global Strategy Files. Unsubscribe in one click, at any time.