TFL: The self-sufficiency illusion

Part one of two

Every year Transport for London announces an operating surplus, and every year the announcement is repeated without inspection. £166m in 2024-25. A third consecutive surplus in 2025-26. The organisation calls this financial sustainability, and the phrase has passed into general circulation the way such phrases do — unexamined, because examining it would require reading a budget, and nobody reads budgets.

The claim is not a lie in the narrow sense. Every figure behind it is published and audited. It is something more durable than a lie: an accounting presentation constructed so that the true statement and the false impression are the same sentence.

What the budget actually contains

The draft TfL Budget for 2026-27 plans revenue expenditure of £9,524.3m, of which £8,971.5m is operating cost, alongside capital expenditure of £2,056.9m. Here is where the money comes from, sorted by the only question that matters: did TfL earn it, or was it taken from somebody by statute?

Earned income of £6,830.3m against revenue expenditure of £9,524.3m is a shortfall of £2.7bn. Against operating costs alone it is £2.1bn. The celebrated surplus exists because £2.5bn of business rates and council tax precept are booked above the line as income. Remove them and TfL runs a structural deficit larger than the entire annual operating cost of the London Underground.

The £1.06bn from the Congestion Charge and ULEZ is a matter of classification, and TfL has classified it in its own favour. These are compulsory charges levied on Londoners who receive nothing in return. Whatever their environmental merits — and they have some — they are taxes. They have been placed on the earning side of the ledger because that is where they flatter.

One further levy does not appear in the table at all, and its absence is instructive. The Crossrail Business Rate Supplement will raise around £290m in 2026-27 from London commercial premises. It is collected by the boroughs for the Greater London Authority, services the GLA's Elizabeth line borrowing — residual debt of some £3.6bn — and never touches TfL's accounts. A compulsory levy funds a railway TfL operates, and TfL's self-sufficiency is computed as though the levy did not exist.

The treatment of that levy also inverts the mechanism described below. BRS revenues rise from April 2026 on the back of the 22.2 per cent average uplift in London rateable values, and unlike retained business rates — where growth may be kept only briefly before a reset — the BRS uplift is permanent. Revaluation growth is stripped out of the money TfL keeps and retained in full in the money that repays the GLA's debt. Whichever pocket a London business pays into, the arrangement has been built to its disadvantage.

How the relabel happened

Two streams of central government money once reached London's transport network by the plainest possible route. TfL received an annual capital investment grant from the Department for Transport, worth around £1bn. The GLA received Revenue Support Grant. Both were unambiguous: sums voted by Parliament, paid to a public body, recorded as grant income. Nobody reading those accounts could mistake whose money it was. Separately, TfL received a general operating grant, withdrawn after the 2015 Spending Review — leaving, on TfL's own description, a funding model heavily dependent on fares, which the pandemic then detonated.

Neither of the remaining grants was abolished. They were renamed.

In 2017-18 the GLA's share of locally retained business rates rose from 20 per cent to 37 per cent. This was not a windfall and it was not devolution. It coincided precisely with the rolling in of TfL's investment grant into the business rates retention system and the withdrawal of the GLA's Revenue Support Grant. The Department for Transport stopped paying. The GLA started paying, out of a tax share transferred for that exact purpose.

No money was created. A grant became a statutory tax entitlement, and the accounting treatment travelled with it. A grant is unambiguously somebody else's money. A retained business rates share can be described, with a little licence and no auditor's objection, as income.

It is worth being precise about how deliberately the system prevents the growth this is sometimes assumed to have delivered. Business rates revaluations are engineered revenue neutral in real terms, and any additional revenue a revaluation generates in London is removed from the capital through adjustments to each authority's tariff and top-up. Even the 100 per cent retention pilot let London keep growth only after revaluation growth had been stripped out. Rising rents in the City and Canary Wharf do not flow to Transport for London. The mechanism excludes them by construction.

Nor is the flow one-way. Since 2016-17 the GLA has paid some £5.9bn of its business rates income back to central government in tariff payments and levy on growth.

What remains is £2,218.2m in the 2026-27 budget — larger than the entire four-year capital settlement, and the biggest line in TfL's accounts after fares. Substantially the same public money the two departments once paid directly, arriving through a different pipe, carrying a story about commercial self-sufficiency that the grants it replaced could never have supported.

And TfL does not dispute where it goes. Its own budget submission states that funding received under the business rates devolution proposals is not restricted to supporting capital investment and may be used to cover operating and financing costs. The £2,218.2m is available to pay wages. TfL says so, in writing, on page 25.

The squeeze was real enough while it lasted. Between 2016 and 2020 capital renewals fell to £388m a year, which is why the Bakerloo line still runs stock built when Harold Wilson was Prime Minister. What the reclassification bought was not more money. It was a different description of the same money, and nobody was appointed to object to the description.

An escalator that answers to nothing

If further evidence were needed that TfL's income is not commercial, it sits in the funding settlement letter. The £2,167m awarded over 2026-27 to 2029-30 was provided, in the Department's own words, against an assumed scenario that fares would rise by RPI plus one per cent in every year of the settlement.

Central government is setting the price of a London travelcard as a condition of a capital grant. The commuter from Warwick or Reading or Sevenoaks pays an annually escalating, centrally mandated charge. TfL banks it as passenger income and reports that it funds itself from fares.

The choice of index compounds the insult. RPI was stripped of its status as a National Statistic in 2013 and is due to be aligned with CPIH in 2030. It runs roughly a percentage point above CPI through a known formula defect. Government uses CPI when it pays money out, on pensions and benefits, and RPI when it takes money in, on rail fares, student loans and vehicle excise duty. The London commuter is indexed to the measure the state has publicly disowned, and then charged a further point on top.

The deeper defect is that the escalator has no relationship whatever to the cost of running a railway. TfL's inflationary cost pressure for 2026-27 is around £228.7m against CPI of 3.6 per cent. Nothing connects that figure to the fare. If costs come in below the escalator, the surplus grows and the passenger has overpaid. If a settlement lands above it, the fare rises on schedule anyway and the shortfall falls on renewals — which is how a sixty-year-old fleet comes to be running under a body reporting three consecutive surpluses.

This is the mechanism's real function. A cost-linked fare would make settlements visible: concede terms above the escalator and something must publicly give. RPI plus one severs the link entirely. Fares rise on their own schedule whatever was agreed in the negotiating room, which means no settlement struck under pressure can ever be shown to be unaffordable. The formula does not transmit efficiency. It suppresses the evidence of its absence.

The only unregulated monopoly

In every other British monopoly, a price rise must be justified. Water companies submit business plans to Ofwat, which challenges their costs line by line, benchmarks them against the most efficient operators and determines allowed revenue for five years. Energy networks face Ofgem under RIIO. Network Rail's access charges are determined by the Office of Rail and Road on identical principles. In each case the firm must evidence its costs, defend its capital programme, and accept an efficiency challenge it did not choose and cannot decline.

Note the sign. The classic British formula, devised by Stephen Littlechild, is RPI minus X: prices rise by inflation less an efficiency factor, forcing real-terms reductions and simulating the competitive pressure a monopolist never feels. Transport for London operates on RPI plus one. The same structure, inverted. Where a regulated utility is compelled to surrender its efficiency gains to customers, TfL is guaranteed real-terms revenue growth with no efficiency test attached to the fare at all.

The justification offered is that democratic accountability substitutes for economic regulation. The Mayor sets fares; Londoners may vote him out. As a substitute it is threadbare. A mayoral election happens every four years, turns on housing and policing as much as transport, and draws a turnout around forty per cent. Ofwat carries a statutory duty to consumers, publishes its cost challenges and can be judicially reviewed. One of these is scrutiny. The other is a manifesto commitment.

The Department does impose efficiency conditions; TfL was required to produce a capital efficiencies plan as a term of the settlement. But the condition attaches to the grant, not to the fare. Whatever savings result are never returned to the passenger. They accrue to the surplus, and the surplus is then presented as evidence of good management.

An organisation with no owner

There is a reason none of this was contested, and it is that nobody holds the job of contesting it.

Transport for London is a statutory corporation created by the Greater London Authority Act 1999. It has no shareholders and no share capital. It issues no equity and cannot be sold. Beneath it sit wholly-owned subsidiaries — London Underground Limited, Rail for London Infrastructure, Docklands Light Railway Limited, Places for London — held through an intermediate company owned by nobody at all. For financial purposes TfL is treated as a local authority, which is why its profit must be called a surplus and reinvested rather than distributed. It borrows against its own balance sheet, and carries debt that no owner stands behind.

Now consider who decides. The Mayor of London chairs the TfL Board. The Mayor appoints every other member of it. The Mayor issues directions under the 1999 Act, sets the level of fares, and approves the budget. And the Mayor is the individual for whom a five-day shutdown of the Underground is a political catastrophe.

Assemble that sequence and the absurdity completes itself. The Mayor appoints the board. The board approves the budget. The budget funds the settlement. The settlement prevents the strike. And the person the strike would damage most is the Mayor who appointed the board. Every chair at the table is occupied by the same interest, negotiating with the only party in the room holding a credible threat. There is no counterparty. Nobody in that building is paid to say the number is too high.

Compare the alternatives. A private monopoly has shareholders who lose money when costs escape and a regulator with a statutory duty to customers. A nationalised industry has a Secretary of State answerable at the despatch box and an accounting officer personally liable for value for money. Transport for London has neither. It has assembled the disadvantages of public ownership — no capital discipline, no competitive pressure, no residual claimant — with the disadvantages of corporate independence, in that no minister is accountable for anything it does.

The London Assembly is the designated check and it is a slight one: it may question the TfL budget but can amend the Mayor's consolidated budget only by a two-thirds majority, a threshold reached approximately never. External audit confirms the figures are correct. It does not ask whether the spending was wise. It is not asked to.

What the critics of this argument would say

Three defences deserve a hearing, and one of them is strong.

First, that the international comparison flatters TfL considerably. Farebox recovery around 60 per cent is exceptional. New York's MTA recovers roughly a quarter of operating costs from fares. Paris depends on the versement mobilité, a payroll tax on employers, for a larger share than TfL takes from business rates. No major urban transit system anywhere covers its costs from fares. Judged against its peers rather than its own press releases, TfL performs well. This is true and it is the strongest thing that can be said.

Second, that the deficit is not where the argument implies. Broken down by mode, the Underground broadly covers itself. The shortfall is the bus network, run deliberately at a loss as a social service for the Londoners least able to afford alternatives. That is a policy choice one may disagree with, but a choice rather than a failure.

Third, that public transport is a public good, that the wider economic return on moving eleven million people is not captured in a farebox, and that demanding cost recovery from a network which exists to make a city function is a category error. There is real force in this. But it is an argument for honest subsidy, openly voted and openly defended. It is not an argument for describing subsidy as commercial income.

What should change

Four things, none requiring a confrontation with anybody.

Report earned income separately. TfL should publish, in its headline results, income generated from services rendered, distinct from compulsory charges and tax transfers, with the operating surplus stated on both bases. This costs nothing and ends the misrepresentation immediately.

Publish pay settlements against a stated affordability test. When a settlement is struck under disruption pressure, the cost, the funding source and the alternative considered should be published. If business rate payers are funding a pay award, they are entitled to see the reasoning. No such entitlement exists today.

Subject fares to a determination. Whether by extending the Office of Rail and Road's remit or establishing an equivalent function with statutory independence, fare increases should require TfL to evidence its costs and accept a benchmarked efficiency challenge, as every comparable monopoly already does. The escalator should carry a negative X, not a positive one.

Attach an automation business case to the capital settlement. If the Department can condition its grant on fare rises of RPI plus one, it can condition it on a costed programme for extending automatic train operation. The alternative is to pay the disruption premium indefinitely and call it a commercial outcome.

 The objection is not that Transport for London spends too much. It may well spend too little; the Bakerloo line is evidence for the prosecution on that charge too. The objection is that a structure has been built in which a substantial public spending decision — indexed, recurring, compounding — is taken by an arm's-length body, funded by a tax nobody voted for, and reported as a commercial result. No minister defends it. No electorate is consulted. The reclassification of 2017 did not increase the money available to Transport for London. It reduced the scrutiny applied to how that money is spent, and that has proved considerably more valuable.

Part two examines what the accounts do with the money once it arrives — and why almost none of the movement in TfL's reported costs was produced by anything TfL did.