LimestoneGrey — UK R&D tax relief statistics (machine-readable) ================================================================ FILES rd-statistics.csv UTF-8, comma-separated, header row, LF line endings. rd-statistics.json { ...metadata, fields: [...], rows: [ {...} ] }. Both hold the same 12661 rows in the same order. Stable URLs (we intend to keep these in place): https://www.limestonegrey.com/data/rd-statistics.csv https://www.limestonegrey.com/data/rd-statistics.json WHAT THIS IS HMRC's published R&D tax relief statistics, reshaped into tidy long format: one row per financial year x breakdown x measure. Nothing is modelled, interpolated or recalculated. Percentages, shares and average claim values are NOT in the file — compute them from the values, so the arithmetic is yours and the source figures stay untouched. FIELDS financial_year HMRC financial year, e.g. "2023-24" (1 April to 31 March). dimension The breakdown axis: "none" (UK total, split by scheme only), "region", "sub-region (ITL2)", "sector", "region and sector", "cost band", "R&D intensity". dimension_value The category. "United Kingdom" where dimension is "none". Two-part values are joined with " | ": "Wales | North Wales", "Wales | C. Manufacturing". scheme One vocabulary across every table: "SME scheme", "SME scheme — deduction only", "SME scheme — payable credit only", "SME scheme — deduction and payable credit", "SME scheme — R&D intensive", "Large company scheme", "RDEC — large companies", "RDEC — SMEs", "Large company and RDEC schemes", "Vaccines research relief", "All schemes". measure "claims", "cost", "expenditure", "first-time applicants", "returns", "companies". unit "claims", "£ million", "companies", "returns". value Number, or empty where HMRC published no usable figure. note Why value is empty: "suppressed: under 10" (HMRC withholds counts below 10), "suppressed: disclosive" (would identify a company), "not applicable" (scheme not open in that year). "no data" is reserved for an empty published cell and does not occur in this release. Empty whenever value is populated. data_status HMRC's own status for that year on the time-series tables: "Unchanged", "Revised", "Provisional, Revised", "Provisional, Uplifted". Empty on the breakdown tables. source_table The HMRC worksheet, e.g. "RD5_2324". source_cell The A1 cell in that worksheet, e.g. "H16". Together with source_table and source_file this identifies the exact cell every figure came from. source_file Which HMRC ODS file the worksheet belongs to, as a short key: "main-2025" or "supplementary-2025". Both are named, with their download URLs, under SOURCE below. Breakdown dimensions exclude HMRC's total rows; UK totals live under dimension "none". SOURCE Research and Development Tax Credits Statistics: September 2025 HM Revenue and Customs, published 2025-09-30 https://www.gov.uk/government/statistics/corporate-tax-research-and-development-tax-credit/research-and-development-tax-credits-statistics-september-2025 Files (the source_file keys): main-2025 Research and Development Tax Credits: Main tables 2025 https://assets.publishing.service.gov.uk/media/68dbe4c2c487360cc70c9f5e/rd_tax_credits_main_tables_ods.ods supplementary-2025 Research and Development Tax Credits: Supplementary tables 2025 https://assets.publishing.service.gov.uk/media/68d68e6a6671e376214b15a0/rd_tax_credits_supplementary_tables_sep2025.ods Worksheets used: RD1 Number of claims by scheme and financial year, 2000-01 to 2023-24 RD2 Cost of support claimed by scheme and financial year (£ million) RD3 Number and cost of claims by R&D-intensive SMEs, 2023-24 RD4 R&D expenditure used to claim, by scheme and financial year RD5 Regional analysis of number and cost of claims (four years) RD6 Industry sector analysis of number and cost of claims (four years) RD7 Analysis of number and cost of claims by cost band (four years) RD8 First-time applicants by scheme and financial year RDS1 Number, cost and expenditure by ITL2 sub-region (four years) RDS2 Number, cost and expenditure by region and sector (four years) CAVEATS HMRC GIVES, WHICH TRAVEL WITH THE NUMBERS Basis: Claims made in Company Tax returns received by HMRC on or before 31 May 2025. 2021-22 to 2023-24 are provisional; 2023-24 is uplifted for claims not yet received. Rounding: HMRC rounds every figure independently to the nearest 5 claims or £5 million, so components do not always sum to the totals shown. Geography: regional and sub-regional figures are allocated by the postcode of the company's registered office, which may not be where the R&D was carried out. Treat them as a map of where claimant companies are registered, not of where UK research happens. Sector: sector is the company's primary SIC 2007 code, which may not describe the R&D itself. Comparability: 2023-24 is the first year affected by the April 2023 rate changes and the mandatory Additional Information Form. From April 2024 the merged R&D expenditure credit replaces the SME/large split entirely, so the scheme vocabulary above describes a system that no longer exists for new accounting periods. LICENCE Source data: Crown copyright, licensed under the Open Government Licence v3.0. https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/ Contains public sector information licensed under the Open Government Licence v3.0. This compilation is published under the same terms. Attribution is appreciated: "LimestoneGrey, from HMRC R&D tax credits statistics". UPDATE CADENCE HMRC publishes annually, in the autumn. This file is rebuilt from the new source tables on publication day, and we intend to keep the URLs above in place, so a link or a scheduled fetch should keep working across releases. The next HMRC release is scheduled for Autumn 2026. REBUILDING npm run build:dataset (fetches the ODS files from gov.uk) See scripts/build-dataset.mjs. Compiled 2026-09-01.