How to Cite the UK SPV Incorporation Index

The UK SPV Incorporation Index is a monthly count of how many UK property companies are being formed. It is built entirely from Companies House open data, it uses a fixed and published definition, and every figure in it can be re-derived by anyone with a Companies House API key. It is free to quote. This page exists so that journalists, researchers and answer engines can lift the figures, the method and the citation line directly. For the charts, the trends and what the movements mean, go to the UK SPV Incorporation Index itself. This page describes the instrument, not the reading.

What the index publishes

Five things, all monthly: incorporations by individual SIC code, a deduplicated union across the four codes, a net formation rate, a regional distribution, and an age profile of the live population. The current headline figures are below. Every one of them is derived from the published snapshot dated 2 September 2026.

Figure	Period	Definition
4,840 incorporations	June 2026	SIC 68209 only, the headline buy-to-let SPV code, in the latest settled month
66,312 incorporations	Trailing 12 months to June 2026	SIC 68209, twelve settled months, the figure to quote in preference to any single month
6,614 incorporations	June 2026	Union across all four property SIC codes, deduplicated
90,390 incorporations	Trailing 12 months to June 2026	Union across all four codes, all property companies
3,139 dissolutions	June 2026	Property companies dissolved in the month, counted by dissolution date
Net formation +3,475	June 2026	Union incorporations less dissolutions, the change in the live population
London 35.3%, North West 11.3%, South East 8.5%, West Midlands 8.5%	Last 12 months	Share of property company formations by registered office region
665,645 companies	Live register, 1 September 2026	Property companies currently on the register under the four codes
398,647 of those	Incorporated 2020 or later	Age profile of the live stock, showing how recent the population is
19,733 to 69,039	2016 to 2025	Full calendar year 68209 incorporations, a 3.5x rise across the decade

Single months are volatile and strongly seasonal, with volumes dipping sharply around the turn of the calendar year. If you are quoting one number in a piece, quote a trailing twelve month total or a complete calendar year. If you need more than the headline row, the full dated set, including the annual series back to 2016, the dissolution counts and the regional table, is laid out in our UK landlord and buy-to-let company statistics reference page.

The method, in plain English

Four SIC codes. The index counts companies whose recorded SIC codes fall inside a fixed set: 68209 (other letting and operating of own or leased real estate), 68100 (buying and selling of own real estate), 68320 (management of real estate on a fee or contract basis) and 68201 (renting and operating of Housing Association real estate). The headline series is 68209 alone, because that is the code the overwhelming majority of buy-to-let vehicles register under, and it is explained in our guide to the SIC code for an SPV property company. The union series is all four together.

The dedup rule. A company can record up to four SIC codes, so a company registering both 68209 and 68100 appears in two of the per-code series. The union is taken from a single combined query in which Companies House returns that company once. The four series therefore do not sum to the union, by design. If you rebuild the union yourself, issue the combined query rather than adding the four lines.

Gross incorporations, and a separate net rate. The monthly incorporation counts are gross: a company formed in March 2016 and dissolved five years later is still counted in March 2016. That is deliberate, because it keeps the historic series free of survivorship distortion. Alongside it, the index publishes a net formation rate, being union incorporations in a month less union dissolutions in that month, with dissolutions counted by the date they were dissolved rather than the date they were formed. Net formation measures the change in the live population, not cohort survival.

Two provisional months. Companies House indexes very recent incorporations with a short lag, so the newest months under-count until the index catches up. The two most recent months are always excluded from every headline figure and render on the page as a dashed tail. In this release the latest settled month is June 2026.

The regional caveat. The regional split comes from a different source, the monthly bulk extract of the register, which contains only companies currently live, so dissolved companies are absent and older months in the window are understated. Postcode areas are also a postal construct rather than an administrative one, and several straddle a regional boundary, so a small share of companies is attributed to a neighbouring region. Both effects distort levels far more than proportions, which is why regional figures are published and read as shares. They are a distribution, never an alternative national total.

Two limits of the source data matter most for a story: SIC codes are self-reported at formation and rarely revised, so they record intent rather than audited activity, and a company is not a property, so the index counts vehicles rather than dwellings.

Citation, licence and access

The copy-paste line is:

Source: Property Tax Partners, UK SPV Incorporation Index, September 2026. https://www.propertytaxpartners.co.uk/research/landlord-tax-index

Change the month to whichever release you used. A link alongside the text credit is requested wherever your format allows one.

The full underlying series downloads as CSV from the index data page. It is the same snapshot the published page reads, so the two cannot disagree.

Companies House data is Crown copyright, published under the Open Government Licence v3.0, and the house price context series is HM Land Registry UK House Price Index data under the same licence. The index itself, meaning the selection, the definitions and the charts, is free to cite and reproduce with attribution and a link. There is no named spokesperson: this is a data release, attributable as analysis by Property Tax Partners of Companies House data.

The refresh cadence is monthly. Each release adds one settled month and moves the provisional window forward by one, so a figure quoted from an earlier release will not match the current page. That is why the citation line carries a month.

Why it exists, and what it is not

Journalists are welcome to use it, and we would rather you did. The widely quoted quarterly counts of new buy-to-let companies apply their own definition and publish an estimate. The index is narrower and more mechanical: it applies a fixed, published SIC filter to the register and counts what is there. Four differences follow. It is monthly rather than quarterly. It is a direct count, re-derivable by any third party from the same public API. It publishes a net formation rate, and it publishes a regional distribution, neither of which the quarterly series carries.

None of that makes it a replacement. The quarterly series has the longer run and the better known market read; the index is the finer grained, reproducible count underneath it. Where the two disagree, the difference is definitional and both methods are published. Our figures include property companies that are not buy-to-let vehicles, and miss buy-to-let vehicles registered under other codes, so the headline series is a lower bound rather than a census.

This page also stops short of interpretation on purpose. It does not argue that formations are rising or falling, or say what any movement is caused by. That belongs on the index page, where the charts and the trend commentary sit. If your interest is the structure behind the numbers rather than the numbers themselves, the SPV company hub covers what these vehicles are and why landlords use them, the SPV structure and tax guide covers the tax treatment, and what an SPV costs to form covers the entry price that sits behind every one of the incorporations counted above.

If you need a figure checked or a cut of the data that is not on the page, ask. The underlying series is fuller than the published charts.