Data & methodology
Every figure on this site traces to a named public source with an as-of date — and this page shows the working. For each data family it lists the primary sources, how often they refresh, how we verify them, and — stated plainly — what the data does not cover. If you are deciding whether to cite us, this is the page to read first; how to cite us is at the end.
The standard every page is held to
Three dates matter on any data page, and we never conflate them: when we pulled the data, what date the source says its data represents, and when the page was last verified. Lagged data presented as current is a lie with a timestamp, so the lag is always part of the fact. Concretely, on every page:
- Primary source per claim. A threshold, date or duty links the law, circular, register or filing it came from — regulator and government sources first; law-firm or vendor material only as a bridge to a primary, never as the source of record.
- Provenance on every table. Tables derived from a dataset carry a visible block: our pull date, the source's own as-of date, the named source, the refresh cadence and the limitations.
- Floors, not false precision. Large machine-derived counts are shown rounded down ("200,000+"), so the stated number stays true between refreshes. Exact figures appear only where a count is small, curated and stable.
- Gaps stated, never implied away. Where a register is partial, lagged, frozen or blocked, the page says so. Anything we could not pin to a primary source sits in a visible "To verify" box, not in the body as fact.
Law & regulation corpus
The reference pages and the clause-level law corpus behind search and ask-the-law cover 85 instruments across the EU, UK, Luxembourg, Ireland, Jersey, Guernsey, Cayman, BVI, Bermuda, the US, Singapore, Hong Kong, Switzerland, Mauritius and the DIFC/ADGM (as at 2026-07-22).
| What | Primary sources | Refresh | How it is verified | Known limitations |
|---|---|---|---|---|
| Full-text law & rulebook corpus (85 instruments) and the per-instrument reference pages | EUR-Lex (EU directives/regulations), ESMA (RTS/guidelines), Legilux and the CSSF (LU), the Central Bank of Ireland and the Irish Statute Book (IE), the FCA Handbook and legislation.gov.uk (UK), the JFSC (JE), the GFSC (GG), CIMA (KY), the BVI FSC, the BMA (BM), the SEC and GovInfo (US), the MAS and Singapore Statutes Online (SG), the SFC (HK), the DFSA and ADGM FSRA (AE), the FSC Mauritius (MU) and Fedlex (CH) | Each source is re-checked for change on its own cycle — frequently-amended rulebooks roughly every 30 days, statutes every 90–180 days — plus on-publication for the domiciles we follow closely. A changed source is re-taken in full, so pages track the law as amended today. | Every page shows a Last verified date. When an instrument moves — an amendment, a new effective date, a new OJ publication — the change is recorded in that page's changelog, dated to when the law moved. | The corpus is the current consolidated position, not a historical archive of prior versions. Five instruments added on 2026-07-21 (three Hong Kong SFC codes, two Mauritius instruments) are not yet under change-monitoring — their first re-check cycle is pending. Short quoted phrases only; the linked official text is always the law of record. |
Fund registers (names & authorisations)
The fund search covers 200,000+ fund names across ten labelled sources. Each pull is a dated snapshot, diffed against the previous one — that is also what powers the new-authorisations view.
| Register | Primary source | Refresh | How it is verified | Known limitations |
|---|---|---|---|---|
| US fund names | SEC (EDGAR) | weekly | Every pull is a dated snapshot; each source is labelled on-page with its own coverage and as-of date, and the search states which layer a name came from. | Registration-based — a name appears when it files, so the list is a floor, not a census. |
| EU fund names | ESMA registers | weekly | Not deduplicated across sources — a Luxembourg or Irish fund appears in both ESMA and its national register. Per-source captures, not a unique-fund universe. | |
| Luxembourg | CSSF | weekly | — | |
| Ireland | CBI register | currently frozen | Served from a static snapshot (2026-07-19) while the refresh path is repaired — stated here rather than hidden. | |
| Bermuda | BMA | weekly | — | |
| BVI | BVI FSC | irregular | Captured 2026-07-19; the source offers no bulk access, so refreshes are manual and less frequent. | |
| Singapore schemes | MAS (CISNet) | ~quarterly | Recognised/authorised schemes, not the full private universe. | |
| Jersey | JFSC | periodic | Fund names are not public — aggregate counts only, no per-fund rows. | |
| Cayman | CIMA | periodic | Aggregate statistics only — the per-fund register is not bulk-accessible. | |
| Guernsey | GFSC | periodic | Partial — aggregate coverage only. | |
US SEC filings — Form D & the ADV provider network
| Dataset | Primary source | Refresh | How it is verified | Known limitations |
|---|---|---|---|---|
| New US private funds (Form D) + the registered investment-company universe | EDGAR full-text search (live filings), the SEC Form D quarterly data sets and the SEC series/class data | weekly (live lane); the historical new-funds layer is re-based quarterly | Offering totals are computed over new notices only — amendments restate a fund's cumulative total, so summing every filing would overcount massively. That rule is applied before anything is published. | Form D under-files — filing happens when the exemption is used, and the SEC's own statistics flag under-capture. US registrations are a floor, not a census. |
| Fund ↔ service-provider network (728,000+ links: auditor, prime broker, custodian, administrator, marketer), on Fund Data | SEC Form ADV Schedule D 7.B.1 — each fund's latest filing, cumulative 2011–2024 | none — permanently frozen at 2024-12-31: the SEC ended fund-level bulk disclosure at that date, so no future refresh of this dataset exists | Built from each fund's most recent filing; the "still active" indicator shown is a derived hint (adviser still SEC-registered, recent filing), labelled as such — never a regulator status. | US-only (SEC disclosure); includes since-closed funds; not cross-linked to the fund-name search (no shared key between the two datasets), so a name in one may not appear in the other. |
Pensions
320+ public pension and superannuation plans, from the plans' own filings and the regulators' statistics — plus the pooled LP-reported private-equity performance lane (2,000+ fund-level records from 5 US plans).
| Dataset | Primary source | Refresh | How it is verified | Known limitations |
|---|---|---|---|---|
| US plans (253): assets, returns, funded ratios | Public Plans Database (Center for Retirement Research at Boston College / MissionSquare Research Institute / NASRA) — used with attribution under its non-commercial terms | re-pulled monthly; the database itself updates as plans' audited reports land | Figures are shown exactly as reported, with the fiscal year stated per plan. | 12–18 month lag — audited reports reach the database 12–18 months after a plan's year end (FY2024 is the current position for 247 of 253 plans). Coverage is the 250+ largest state/local plans; corporate US plans publish no comparable returns. |
| Australian super funds (77): fund-level statistics | APRA annual fund-level superannuation statistics (June 2025 edition) | per APRA edition (annual), re-checked monthly | Spot-checked against funds' own published figures. | APRA-regulated funds only; the workbooks carry APRA's own copyright — cite APRA when reusing these figures. |
| LP-reported private-equity performance (5 plans, 2,000+ fund records) and the cross-plan divergence view | Each plan's own published PE report: CalPERS (as of 2025-09-30) · CalSTRS (2025-06-30) · Washington WSIB (2025-12-31) · Oregon OPERF (2026-03-31) · Minnesota SBI (2026-03-31) | re-checked monthly; each plan publishes quarterly to annually | Each lane's summed commitments are reconciled against the source document's own totals row before publication — a lane that does not reconcile is not published. Fund matching across plans is conservative: unmatched stays unmatched. | These are one LP's numbers, not fund-wide returns. Plans report to different as-of dates and on their own bases, so part of any cross-plan gap is timing. Oregon flags funds sold in the secondary market; those are carried with that caveat and excluded from divergence statistics. |
Pension figures as reported by each source · US plan data:
Public Plans Database (Center for Retirement
Research at Boston College / MissionSquare Research Institute / NASRA), used with attribution
under its non-commercial terms · AU data: APRA ·
PE data: each plan's own published report · Limitations: lag and coverage as stated in the
table above.
Calendar
The calendar is authored from primary sources, not scraped: a structured register of 110 dated obligations (as at 2026-07-19) — 41 regulatory, 50 tax, 12 investor-reporting, 4 ESG and 3 financial-statements — each row carrying its own source link and last-verified date.
| What | Primary sources | Refresh | How it is verified | Known limitations |
|---|---|---|---|---|
| Regulatory, tax and reporting deadlines across the covered domiciles | 10 regulators (CSSF, CBI, FCA, JFSC, GFSC, SEC, ESMA-derived EU law, MAS, DFSA, ADGM FSRA) and 13 tax authorities — IRS, HMRC, Revenue.ie, Luxembourg ACD/Guichet.lu, the Jersey and Guernsey revenue services, IRAS, BZSt, Japan's NTA, Korea's NTS, the UAE FTA, OeKB and the Swiss ESTV | monthly verified sweep of every row, plus same-day updates on deadline news | Of the 110 rows, 91 rest directly on a primary source; 12 rest on named secondary guidance (marked as secondary on the row) pending a primary; 7 are marked uncertain and are flagged as such, not presented as settled dates. | Deadlines are shown as stated by the authority — where a date depends on a fund's own year end, the rule is given rather than a guessed date. Public-holiday data (9 jurisdictions, 2026–2029, from a maintained public-holiday dataset) is used to roll only US day-precision deadlines; for all other jurisdictions holidays are display-only and stated deadlines are never shifted — the conservative error. |
Corrections policy
The record tracks the rules, not our housekeeping — and it is never silently rewritten:
- Last verified, on every page. The date under the title is when the page was last checked against its sources. If a page looks stale, the date will tell you.
- Changes logged where you read them. Law and calendar pages record substantive changes — amendments, new effective dates, threshold moves — in a dated changelog on the page itself, dated to when the rule moved. A machine-readable feed (RSS) aggregates these entries.
- Corrections are content. When we get a fact wrong, it is corrected in place and the changelog entry states the correction — what was wrong and what is right. Superseded entries are never deleted.
- "To verify" means exactly that. Anything not yet pinned to a primary source sits in a visible box, stated as open — never asserted as fact.
- Anyone can trigger a correction. Prove a page wrong via the contact form and the fix becomes a same-day changelog entry.
How to cite us
Cite the page, its as-of date, and — where you can — the primary source next to it. Every claim on this site links its primary source precisely so that citing us never has to mean trusting us.
- Format: FundRegTracker, "[page title]", fundregtracker.com/[page], as at [the page's Last verified date]. For dataset figures, carry the source's own as-of date from the provenance block — it is part of the fact.
- Reuse: aggregates and figures we derive from public registers, filings and law texts are freely reusable with credit (a link back to the page).
- Exception — PPD-derived figures: anything derived from the Public Plans Database (the US pension plan figures) is subject to PPD's non-commercial terms — credit the Public Plans Database (CRR / MissionSquare Research Institute / NASRA), and do not reuse those figures commercially without permission from PPD.
- APRA-derived figures (Australian super) should credit APRA as the source, per the copyright notice on its statistics.
The gotcha, for anyone reusing this data: the most common citation error is not a wrong number — it is a right number presented as current when its as-of date says otherwise. A FY2024 funded ratio or a 2024-frozen provider link is accurate for its date. Keep the date attached and the citation stays honest.
To verify
- Public-holiday lists — each jurisdiction's calendar was spot-checked on distinctive days (e.g. Liberation Day for Jersey/Guernsey, Juneteenth for the US), but the full 2026–2029 lists have not been verified line-by-line against official government gazettes.