The Subsidy Clock

What renewable electricity support costs Britain

Methodology & sources

Every figure on the counter links to its source. Most are official statistics; the exceptions are listed under Sources that are not official statistics. This page sets out, scheme by scheme, where the data comes from, how the numbers are computed, how often they update, and what the known limitations are.

On this page: Perspectives · Indirect costs · How the headline figure is made up · Nominal & 2024 prices · Sources that are not official statistics · Equivalences · Map & asset panels · Data freshness and the running total · Corrections · Citing this site

Contracts for Difference (CfD)

Source: LCCC, Actual CfD Generation and Avoided GHG Emissions · cadence: daily · CfD – renewables, including biomass; nuclear is tracked separately and appears once Hinkley Point C generates

Formula in words: the Low Carbon Contracts Company publishes the daily payment made under each individual CfD contract. We sum these per-contract daily payments. When the market reference price exceeds a contract’s strike price the daily payment is negative — the generator pays back — and these negative payments are netted against positive ones. Both figures are computed: the gross total (positive payments only) and the net total (after paybacks). The counter and all headline figures use the net total, which is the lower of the two.

The recipient table covers the CfD schemes (from LCCC per-generator payments) and the Renewables Obligation (named generating stations from David Turver’s ROC reconstruction, valued at buy-out); the other schemes are held as totals in version 1. CfD rows group a station’s separate per-phase contracts into one physical asset, expandable to the individual contracts. RO per-station figures are buy-out value only — the directly-sourced per-generator basis — so they understate the full cost by the recycle element that the RO scheme total includes; the bottom-up RO buy-out reconstruction (≈£59.9bn) reconciles with our recycle-inclusive scheme total (£74.6bn), the difference being recycle value plus the 2002–07 years the reconstruction does not cover.

Constraint payments (paid to switch off)

Source: Elexon BMRS, balancing settlement bid stacks · cadence: daily · history: REF published annual totals

Formula in words: for each settlement period, the accepted balancing-mechanism bids of wind units are summed: the cost is the bid volume multiplied by the bid price, and the curtailed volume is the energy the unit was instructed not to generate.

All accepted bids from wind BM units are counted, regardless of soFlag. The soFlag governs imbalance-price tagging, not whether the unit is paid. Constraint costs for 2010 to 2023 use the Renewable Energy Foundation’s published annual totals.

Constraint costs for 2010 to 2023 come from the Renewable Energy Foundation’s published annual constraint totals; from 1 January 2024 they are measured bottom-up from Elexon settlement data. Bilateral trades agreed outside the balancing mechanism are excluded, so the constraint figures are a lower bound on what is actually paid for curtailment.

Renewables Obligation (RO)

Formula in words: annual official scheme values from Ofgem’s Renewables Obligation annual reports — the number of ROCs presented multiplied by the worth of a ROC, or the headline scheme value where Ofgem states it directly, unrounded. Mutualisation is added in the scheme year it makes good: when suppliers failed in 2017–2021 without paying what they owed the scheme, the remaining suppliers covered the shortfall (£525.5m received). In 2024 prices an April–March scheme year is priced three quarters at its first year’s CPIH and one quarter at the next.

Feed-in Tariffs (FIT)

Formula in words: what licensed suppliers paid into the Feed-in Tariff levelisation fund each scheme year (April–March), from Ofgem’s annual reports and annual levelisation notices: the money that reached bills, as Reliable Britain counts it. In 2024 prices each scheme year is priced as the Renewables Obligation’s.

Perspectives: renewables, and all low-carbon

The dashboard's headline counts renewable generators — wind, solar, hydro and the biomass plants (Drax and the like) that hold renewable contracts and count towards the UK's renewable targets. The all-low-carbon view adds the one kind of low-carbon generation that is not renewable: nuclear. Hinkley Point C has not yet generated, so its CfD payments stand at : the all-low-carbon total () is for now identical to the renewables headline (), and the gap will open only once Hinkley begins to pay out. Indirect costs (network, balancing, carbon pricing, capacity) are tracked separately and always marked estimated. The dashboard's Include indirect costs switch controls whether they are added to the headline counter; it starts on, so the figure you first see is the combined total, and the counter's own caption says so. Switch it off and the counter falls back to the direct, measured payments alone.

Indirect costs (estimated)

Beyond the direct schemes above, renewable generation imposes costs that reach bill-payers and taxpayers through network charges, balancing costs, carbon pricing and the capacity levy. The dashboard shows these as ‘the indirect bill’, always marked estimated: each component’s amount is measured from official and primary sources (Ofgem, NESO, HMRC, LCCC, National Grid, the NAO and ONS), or modelled by the engine behind Reliable Britain where no measurement exists, as each component’s source notes say, and the share attributable to net-zero policy is a stated judgement. The attribution rule for each component is quoted verbatim below. From 2017 onwards the BSUoS series is measured from NESO’s settled data; the hand-maintained years end at 2016 because the settled data begins on 1 April 2016 and so covers no complete calendar year before 2017.

Capacity Market

The Capacity Market pays power stations, batteries, interconnectors and demand response to be available at times of system stress; bill-payers fund it through a levy. Any power system needs backup for the coldest, stillest evening, so the whole levy is not a cost of net zero. Since 13 October 2026 the Clock counts only the part net zero caused, the same rule it applies to network and balancing charges, which count only the cost above what a Britain without net-zero policy would have paid.

What is counted. For each year: the Capacity Market actually paid (LCCC), less what the same system with net-zero policy removed would have paid for its own backup at that year’s capacity price. That system keeps the actual demand, nuclear, interconnectors and pumped storage; builds no wind or solar; keeps coal open where it pays; and builds gas to meet the remaining peak. Its own capacity bill is the actual price per kilowatt times the firm capacity it would have contracted. New and kept-open plant in that system are already charged at their full build and keep-open cost in the model, so they are not charged a capacity payment as well. The resulting share is applied to each month’s payment in that year.

Why the figure is a floor. The system without net-zero policy is charged the capacity price that actually cleared. Since 2021 a 550 g/kWh emissions limit has kept the 11 GW of coal that bid into earlier four-year-ahead auctions out of them (Ofgem, Annual Report on the Operation of the Capacity Market 2020/21 and 2021/22), and clearing prices rose after it. Without net zero that coal would have bid and held prices down. No published source measures how much of the price rise net zero caused, so none of it is counted.

Range. Low: kept-open coal in the system without net zero would also have needed a capacity agreement. High: the whole levy, the rule the Clock used before 13 October 2026. Both are shown below. The shares for 2016–2025 come from the model of the system without net-zero policy behind Reliable Britain (Report No. 1; engine release v4-s43). 2026 onward uses the 2025 share until the model is run for those years. The per-year shares, the amounts and their provenance are published in breakdown.json (netzero_share) and summarised on the change log. Attribution confidence stays medium.

Climate Change Levy

Only electricity’s share of main-rate CCL declarations is counted, as HMRC receives it; gas and solid-fuel CCL is excluded. Carbon Price Support, the tax generators pay on their fuel, is counted in the carbon line below. The figure errs low.

Carbon in the electricity price

Generators that burn gas or coal pay for the carbon they emit: an allowance for every tonne of CO₂ and, from 1 April 2013, Carbon Price Support on their fuel. The most expensive plant running sets the price of power, so the carbon cost raises the price of all electricity sold. This line counts that whole effect on the wholesale price, year by year, as the engine behind Reliable Britain models it (Great Britain).

It is shown in two parts. Tax collected by the Treasury: Carbon Price Support (HMRC receipts) and the power sector’s share of emissions-trading auction receipts; for 2013–2020 that share is estimated, because no publication attributes auction revenue to sectors. The rest of the carbon in the price: the rise in the price of power that no one collected as tax, which generators kept. From 2005 to 2012 generators were given most of their allowances free and paid little for them, but the carbon price still set the price of power, so those years count, as in Reliable Britain. Attribution confidence is medium.

VAT on household bills

Households pay 5% VAT on their electricity bills, so the counted costs that reach household bills carry VAT too. This line is 5% of households’ share of each counted line, as Reliable Britain counts it; households’ share of electricity each year comes from the same engine. The Climate Change Levy is left out, because households do not pay it.

Transmission charges (TNUoS)

Each year’s charge is compared with what the same year would have cost in a Britain without net-zero policy: the 2002–05 grid carried forward, as the engine behind Reliable Britain models it. Only the difference is counted, and a year below that comparison counts as negative. The part of the charge that generators pay is left out of both, because generators recover it through prices and subsidies already counted.

Balancing costs (BSUoS)

As for TNUoS: each year is compared with what balancing would have cost the same year in a Britain without net-zero policy, as the engine behind Reliable Britain models it; only the difference counts, a year below it counts as negative, and the generators’ part of the charge is left out of both. For the current year the comparison is cut to the days the data cover.

Corrected 30 September 2026. This series previously came from NESO’s Daily Balancing Costs dataset, which breaks down balancing-mechanism and ancillary spend. That is not the same quantity as the balancing cost recovered from bill-payers through BSUoS: it ran 23–39% below the outturn in NESO’s own Annual Balancing Costs Report in every year from 2018-19. Because the pre-2017 history and the baseline are both on the gross BSUoS basis, the old series joined two different definitions mid-run. The series now reads NESO’s settled BSUoS charge data — the half-hourly charge up to 31 March 2023, and actual BSUoS cost thereafter, because the April 2023 fixed-tariff reform broke the identity between what users are charged and what balancing costs. Full detail, and the effect on the totals, is on the corrections page.

Paid to switch on (constraint turn-up)

When wind farms are paid to switch off behind a network bottleneck, the grid operator also pays other stations, mostly gas, to switch on and replace the lost energy. That cost is recovered through BSUoS, so it was already inside the balancing uplift. It is now taken out of the uplift and shown as its own line; the balancing line falls by the same amount and no total changes.

The figure is an estimate, built daily from Elexon settlement data for every half-hour since November 2015, using the method of Octopus Energy's Wasted Wind tracker. In each half-hour, the volume of wind bids accepted as constraint actions is priced against that half-hour's accepted offers: offers not marked as constraint actions first, then in the order they were accepted, with a part-share of the offer that crosses the volume. Short-duration offers are left out, as Octopus leaves them out. Together with the payments to switch wind off, our figures reproduce Octopus's published total to within 0.1% (2025: £1,465.7m against £1,466.8m); turn-up alone came to £1,083.6m (Octopus: £1,083.7m). It is counted with the indirect costs, not the direct payments, because the money goes to the replacement stations, not to renewable generators, and because it is an estimate rather than a record of payments.

As a cross-check we also total what was actually paid on offers marked as constraint actions, from stations other than wind farms, in the half-hours when wind was being switched off. That figure is lower than the estimate, because much of the replacement energy is bought through offers not marked as constraint actions; it is a floor, not a contradiction. NESO's own thermal-constraint cost is higher than the estimate (2025: £1,831m), because it covers every thermal constraint, not only replacing switched-off wind. Conservative direction: the line can never exceed the balancing uplift it is taken from.

The double-counting guard

Constraint payments to wind farms are counted in the direct layer (‘paid to switch off’). Those same payments are recovered from bill-payers through BSUoS, so the BSUoS series is netted of direct constraint payments before attribution — the same pound is never counted twice across the direct and indirect layers.

Indirect costs not counted

How the headline figure is made up

Price basis for real-terms figures: CPIH, 2024 — see Real terms; annual averages, and ONS’s published months for a year not yet averaged; today’s money is priced at the latest month.

Share of the electricity bill

The "Subsidy as a share of the electricity bill" chart expresses the renewable-energy subsidy as a fraction of total UK electricity consumer expenditure. The denominator is DUKES 1.3, Sales of electricity and gas by sector (annual, calendar-year).

The numerator is this site's own subsidy total, which is a deliberately conservative lower bound, so the share is a lower bound too. The 2022–23 energy-price spike raised total electricity expenditure and so lowers the subsidy share in those years even though the subsidy itself kept rising.

The chart, and the share-of-bill figure under the counter, use complete years only. A year is left out until every scheme in it has reported: the Renewables Obligation, Feed-in Tariffs, Climate Change Levy, carbon pricing and TNUoS publish a year or more in arrears, so a newer year would set a part-year of subsidy against a full year’s bill. The year shown is named beside each figure; at present it is 2024.

Real terms (2024 £)

Reconciliation against official totals

The bottom-up CfD series (per-contract daily payments) is checked against the LCCC’s official In-period Tracking daily totals, summed over settlement dates present in both series. The most recent days are excluded to allow for the settlement lag.

MonthDays comparedBottom-upOfficialDivergence%

What is not included

The totals deliberately exclude several subsidy mechanisms, so they are conservative:

Sources that are not official statistics

Most figures come straight from official statistics (LCCC, Ofgem, Elexon, NESO, DESNZ, ONS). These do not:

The recipients map and asset panels

The map shows one marker per station per scheme (a plant paid under both CfD and RO appears once per scheme), positioned by the station’s coordinates, with marker area proportional to cumulative payment. Clicking a marker opens that asset’s panel. Definitions used in the panel:

Basemap: OpenFreeMap (© OpenMapTiles, data from OpenStreetMap). Terrain: Terrain Tiles on AWS (Mapzen/Tilezen Joerd), which incorporates among others Copernicus EU-DEM (produced using Copernicus data and information funded by the European Union), United Kingdom terrain data © Environment Agency copyright and/or database right 2015, and USGS SRTM/GMTED2010 data — full attribution list. Map tiles are fetched by your browser directly from those providers; no identifier beyond the tile request itself is sent.

2024 prices, and cash

The headline is in 2024 prices: each year's cost converted to 2024 prices with the ONS CPIH index (2024 = 100), so that a pound paid in 2002 and a pound paid in 2026 are worth the same. The same annual series is shown here in cash, money summed as it was actually paid. The cash total is smaller, because early-years subsidy was worth more in 2024 prices.

YearNominal2024 prices

The equivalence figures

The dashboard's comparison lines use deliberately conservative denominators, and every count divides the quoted £280bn+ floor — not the raw total — then rounds down:

Data freshness and the running total

Every source publishes in arrears, each to its own date. The counter is the published figure for each line up to that line’s latest data, plus an estimate since then: each line carried forward from its own data end at its latest yearly rate. That rate is the last complete year for annual lines (RO, FiT, the levy, carbon, grid charges), the trailing 365 days for daily feeds (CfD, constraints, balancing, turn-up), and the last 12 months for the Capacity Market; VAT is carried with the lines it is charged on. The estimate is shown separately on the dashboard (‘Estimated since the latest data’) and is replaced by published data as each source updates. In 2024 prices, the estimate is priced at ONS’s latest monthly CPIH.

Each scheme on the dashboard carries a badge showing its update cadence and the latest date its data covers. Every source publishes in arrears — LCCC releases daily CfD data weekly, NESO publishes settled BSUoS data roughly three weeks behind the days it covers, and Capacity Market monthly data follows the settlement calendar — so a publication lag is normal and is not flagged. A scheme is marked stale only when its latest data is older than that scheme’s normal publication lag: 21 days for CfD, 3 days for constraints, 30 days for BSUoS, 75 days after the end of the latest data month for the Capacity Market, and 2 years for RO and FIT annual reports.

The BSUoS threshold was 7 days until 30 September 2026, when the series moved to NESO’s settled BSUoS data. That source is slower than the daily extract it replaced — about three weeks behind rather than one — because settlement runs take time to finalise. The Clock accepts roughly twelve days less timeliness for a figure that measures the right quantity; see corrections.

Restatements

Source datasets are revised by their publishers. When a source revises history, the change is logged as a restatement; figures are never silently overwritten.

That log covers the sources this site fetches automatically. It distinguishes a publisher restating its own history from a change of basis made here: the BSUoS entries dated 30 September 2026 are the latter, and are labelled as ours with the cause named. Several series are maintained by hand from annual publications, and those are revised too. On 1 September 2026 the electricity spending denominator was moved to the DUKES 2026 edition, published 30 July 2026, which revised its own history from 2010 onwards: 2024 consumer spending on electricity rose from £65.58bn to £66.19bn (+0.94%), 2023 rose £686m, and 2022 fell £222m. The figures previously shown were a correct reading of the 2025 edition, so this is a restatement by the publisher rather than an error here. DESNZ has said it plans a methodology review of this table, so further revision is likely in the 2027 edition.

Because publishers restate, a figure is only meaningful alongside the edition it came from, and this site now records that edition with each hand-maintained figure. Two related cautions apply to anyone checking the workings. Government asset URLs are not stable citations: the address that served the 2025 DUKES workbook now redirects to the 2026 file, so re-fetching it to verify an older figure silently returns different data. Superseded editions are therefore cited here through the UK Government Web Archive.

Corrections

This site reconstructs official data, and reconstruction can contain errors. If you find one, I want to know. Confirmed errors are corrected and logged on the corrections page, with credit to the finder if they wish. Report an error.

Citing this site

Every dashboard card has a stable anchor (e.g. subsidyclock.co.uk/#switch-off) — cite any figure using its anchor link and the formats below. To cite the site as a whole:

Anchors are a public contract: they will not be renamed. Data downloads and their licence are on the data page.

Figures computed . Questions about the data should start with the source links above.