Transport for London says fare evasion costs it about £190 million a year and ran at 3.5 per cent of journeys in 2025-26. The figures underpin a strategy to halve evasion to 1.5 per cent or less by 2030-31 and are reported to the TfL board to a single decimal place. In March the commissioner told the board the rate had fallen by 0.1 points in a quarter. In June TfL said its approach was working.
Tube Alerter asked TfL for the methodology behind the estimate under freedom of information law. Its internal review disclosed that the estimated value includes a "10% reduction" reflecting "medium confidence" in the assumptions behind the calculation, and that TfL has no system matching passenger journeys to payments. Later correspondence described the same adjustment as a "10% margin of error". A reduction lowers a figure. A margin of error puts a range around it.
Asked which was correct, TfL said: "The estimated value of fare evasion includes a 10 per cent adjustment to reflect the limitations of data used to produce a network-wide estimate ... The adjustment is applied to ensure the reported figure appropriately reflects an element of uncertainty within the methodology and does not overstate the estimated level of revenue loss."
The £190 million is therefore the figure after the adjustment. TfL declined to say what it is before. On its own description, the unadjusted estimate is about £211 million.
TfL has not published an equivalent adjustment or uncertainty range for the rate, which is drawn from the same surveys, enforcement data and assumptions. The 1.5 per cent target, the quarterly figures reported to the board and the comparisons with other cities all rest on that percentage.
Asked where the adjustment had been disclosed, TfL said:
TfL's own documents treat uncertainty differently elsewhere. The quarterly report that carries the fare evasion table records that road safety figures are factored to allow for late reporting, that bus crime data is "over 90 per cent accurate" and that crime statistics follow the code of practice for statistics. Fare evasion appears with no equivalent note. The safety and security panel was told the rate averaged 3.4 per cent across the first three quarters of 2024-25 against a target of below 3.3 per cent, a miss of a tenth of a point.
The bus survey
TfL also confirmed how the rate is produced on the bus network. Buses carry about half of all TfL journeys and have no gatelines, so passenger surveys form a central part of the estimate.
The survey covers 180 sampled duties a quarter, unweighted, with the rate calculated as irregularities divided by total interviews. Asked how passengers who are missed or who refuse are treated, TfL said:
A passenger without a valid fare has a reason to decline. Those passengers leave the survey calculation rather than counting against it, so the risk is of undercounting rather than overcounting, and it runs against the direction of the 10 per cent reduction. TfL declined to provide interview totals or refusal rates, leaving its size impossible to establish from outside.
TfL says the survey is one input among several and its answer is scoped to the survey irregularity rate. On that account, the largest mode on the network is measured by combining a survey that evaders can opt out of with enforcement data TfL says is targeted rather than representative. The output is reported to a tenth of a point: bus rates of 2.4, 2.4 and 2.9 per cent across three quarters of 2024-25, a rise of half a percentage point, or about a fifth in relative terms, in a single quarter.
Where the figures are used
The estimate is used well beyond TfL's own reporting. In September last year the assembly's transport committee pressed TfL on the target, citing lost revenue of almost £190 million and polling suggesting about 80 per cent of Londoners think TfL is not doing enough. City Hall Conservatives have campaigned on the same numbers. When the mayor approved raising the penalty fare to £100, the mayoral decision recorded TfL's estimate of more than £130 million a year.
The figures are used to reassure as well. TfL's April 2025 strategy release noted New York's metro rate of 13 per cent and said the same rate in London "would equate to over £400m of lost revenue per year".
The step from a percentage of journeys to a figure in pounds cannot be reproduced from anything TfL has published. Its rate counts journeys rather than income: 3.5 per cent of about 3.6 billion journeys, or roughly 126 million unpaid. Turning that into a cash figure requires an assumption about what each evaded journey would have earned, which TfL does not disclose. The consequence shows in its own published figures. In April last year TfL put 2023-24 at 3.9 per cent, "which equates to a revenue loss of more than £130m". For 2024-25 it reports 3.5 per cent and about £190 million. That is fewer evaded journeys, with a considerably larger sum attached to them. A change in method between the two years, or in the mix of modes and journey lengths, could account for some or all of the difference, and TfL described the earlier figure as a minimum. Establishing which would require the conversion assumption, a second lever on the published figure sitting on top of the 10 per cent.
TfL's published figures for earlier years vary between documents. Its April 2025 press release gave 2023-24 as 3.8 per cent. The strategy document published in the same month gave 2023-24 as 3.9 per cent. By June this year TfL was attributing that 3.9 per cent to 2022-23 instead. The £130 million estimate was published alongside the 3.9 per cent figure.
The assembly has met the same response. Asked for fare evasion data before its July 2025 meeting, TfL set out its method in one paragraph listing gateline, ticketing, enforcement and survey sources, with no reference to the adjustment or the confidence assessment. It refused breakdowns by station and by barrier type because disclosure "would be of value to fare evaders". The same document shows how much of the framework is assumed: TfL told the committee it has no exact split of evader types but works on 5 per cent accidental, 40 per cent opportunistic, 30 per cent calculated and 25 per cent chronic.
Measuring the trend
TfL's use of enforcement output in the estimate has been public for years. The internal review adds the mechanism: contactless rates identified through revenue inspection charges, bus rates resting on deployments and operational activity, and Underground rates combining charge rates with gateline data and output from TfL's Irregular Travel Analysis Platform into what TfL calls a "balanced rate".
Enforcement targeted at hotspots can overstate the level of evasion if it is not adjusted for, and TfL cites this in explaining why it discounts the final figure.
The target is a level, but progress towards it is judged on the movement in the rate. In 2025-26 TfL raised checks from 4.6 million to 6.9 million and increased the revenue inspection charge. Both alter what the enforcement input records without passenger behaviour necessarily shifting. TfL says as much, adding that revenue recovered is not a like-for-like comparison between years. The figure that would allow one, the number of charges issued, was requested and withheld.
How others measure it
TfL benchmarks itself against New York, whose rate it cited last April. New York's method has been through a public review that TfL's has not. The MTA's inspector general reviewed six quarterly fare evasion surveys, found the data collection flawed and worked with New York City Transit and a Columbia University statistician to redesign the method. The redesign was published. Bus surveys are stratified by borough, with each stratum weighted by an estimate of total ridership, paid and unpaid. Bus evasion is now calculated using automated passenger counters and released as a public dataset, which bodies such as the Citizens Budget Commission use in their own estimates of the cost. TfL's bus survey applies no weighting, and passengers who decline an interview leave the sample.
New York's method has itself been criticised and revised, and the Citizens Budget Commission has said the MTA does not provide consistent and complete data on commuter rail and toll evasion. Closer to home, the Department for Transport estimates evasion of vehicle excise duty, commissioned the University of Southampton to review the methodology behind it, and publishes 95 per cent confidence intervals for its most recent estimates where possible.
An estimate that cannot be checked
The disclosed features of the method point in opposite directions. Survey non-response carries a risk of undercounting. Targeted enforcement can overstate, which TfL gives as its example of why it discounts. The only adjustment TfL has disclosed is a flat 10 per cent for the limitations of the data overall, derived from an assessment it will not publish, and TfL has not said which limitations it covers. The published material does not allow the correction to be tested, either by the board members and assembly members asked to judge performance against the target or by anyone else.
TfL's internal review explains the refusal. The withheld material includes inspection data, anticipated methods of fare evasion and vulnerabilities, which TfL says are baked into the calculations. Its press office puts it more broadly, saying publication "could compromise TfL's ability to prevent, detect and respond to fare evasion". On that reading, the workings cannot be released at any point.
New York's inspector general took the opposite view when it opened up the MTA's method, saying that the public deserved the best possible methodology.
TfL declined to give the unadjusted estimate, the number of revenue inspection charges issued in 2024-25 and 2025-26, interview totals and refusal rates for the bus survey, the periods covered by its 3.5 and 3.3 per cent figures, or whether it regards counts and totals of this kind as inseparable from operationally sensitive material.
TfL's response
A TfL spokesman said: "Robust analysis of fare evasion on London's transport network is a vital tool in tackling it, and the estimated total cost of fare evasion is just one of many figures we produce to measure the scale of the issue. All figures produced are estimates that are calculated according to TfL's best interpretation of the data, in line with industry standards. The estimated cost of £190m represents an evidence-based assessment that takes limitations of the data used into account."
TfL said its estimate and confidence assessment were developed after the pandemic by its security, policing and enforcement and strategic problem-solving teams, that the methodology and assumptions are held in internal working documents, and that its public reporting focuses on performance and estimated financial impact rather than the technical methodology behind it. It said it was committed to transparency and published information that provides confidence in the estimated figures.
Sources & methodology
This article is based on a freedom of information request made by Tube Alerter to Transport for London for the methodology behind its fare evasion estimate, TfL's response, TfL's internal review of that response, and subsequent correspondence with TfL's press office in which the quotations attributed to TfL in this article were given. Quotations from TfL are reproduced verbatim from that correspondence and from the internal review; where an ellipsis appears in a quotation, intervening text has been omitted without changing its sense. Published figures are taken from TfL's April 2025 fare evasion strategy and the press release issued alongside it, TfL board papers and safety and security panel reports, and TfL's written material for the London Assembly transport committee's July 2025 meeting. One figure in this article is derived rather than published: the unadjusted estimate of about £211 million is £190 million with TfL's stated 10 per cent reduction reversed, set out here so it can be checked. TfL declined to confirm or give the unadjusted figure. Where TfL has published more than one rate for the same year, both are given rather than one being selected. Comparisons with New York rely on the MTA inspector general's published review of New York City Transit's fare evasion surveys, the revised methodology issued after it, and the Citizens Budget Commission's published work on fare and toll evasion; the domestic comparison relies on the Department for Transport's vehicle excise duty evasion statistics and their accompanying notes and definitions. TfL was given the substance of each finding in this article and the opportunity to respond, and its response is published in full above. Its declines are listed rather than characterised.
- TfL freedom of information - the request, the response and the internal review disclosing the "10% reduction", the "medium confidence" assessment and the absence of a system matching journeys to payments.
- TfL board and panel papers - quarterly safety, health and environment reports carrying the fare evasion table, and safety and security panel papers reporting the rate against target.
- TfL press releases - the April 2025 fare evasion strategy announcement, including the New York comparison and the "over £400m" figure.
- London Assembly transport committee - correspondence with TfL on fare evasion and the written material provided for its July 2025 meeting, including the assumed split of evader types.
- Greater London Authority decisions database - the mayoral decision approving the £100 penalty fare, recording TfL's estimate of more than £130 million a year.
- MTA Office of the Inspector General, fare evasion review - the review of six quarterly surveys, the findings on data collection, and the redesign carried out with New York City Transit and a Columbia University statistician.
- Citizens Budget Commission - independent estimates of the cost of fare and toll evasion in New York, and its assessment of the MTA's commuter rail and toll data.
- New York State open data - MTA bus fare evasion figures released as a public dataset.
- Department for Transport, vehicle excise duty evasion statistics - roadside survey estimates published with 95 per cent confidence intervals, and the accompanying notes and definitions setting out the methodology and its independent review.
- TfL publications and reports - annual journey totals used to convert the published rate into a number of evaded journeys.
We build highly customisable tube alerts for the London Underground, Overground, DLR and Elizabeth line. Choose your lines, set your time window, pick your days, and filter by severity. We will only email you when something actually affects your commute.
Set up a free alert →
Good Service, 882 Cancelled Trains
Six weeks of National Rail data against TfL's own status feed on the Mildmay line.
TfL Fares, Subsidy and the Funding Chart
The chart everyone posts to explain London's fares, and what it actually shows.
21 Years of Tube and Bus Fares
Every TfL fare change from 2005 to 2026, with inflation adjustment.
We will update this article if TfL releases the unadjusted estimate, the confidence assessment or the bus survey response rates. Send us a tip if you work on fare evasion measurement and can add context we are missing.