Topic

Pensions Dashboards

Pensions dashboards are an electronic communications service that will let individuals view their pensions information — including the State Pension — for free in one place online, and be reunited with lost or forgotten pensions ^DWP. The programme is now in its connection phase: roughly 2,600 schemes are required to connect, with the legal deadline of 31 October 2026 approaching and the public MoneyHelper dashboard “on the horizon” ^TPR. Connection and launch are separate events: dashboards are not yet available to the public, and MaPS expects the MoneyHelper dashboard to be available in 2027/28 ^HoC.

Legal framework

The duties sit on a chain of primary then secondary legislation. Part 4 of the Pension Schemes Act 2021 inserted sections 238D–238G of the Pensions Act 2004 to require occupational schemes to connect to the dashboards ecosystem; section 121 placed corresponding obligations on the FCA for personal and stakeholder providers (now COBS 19.11 of the FCA Handbook) ^DWP.

The Pensions Dashboards Regulations 2022 (SI 2022/1220, in force 12 December 2022) set the operational detail: which schemes are in scope, the connection and data duties, and the “find / view” request mechanics ^leg. Parts 3 and 4 apply to trustees or managers of relevant occupational schemes with 100 or more relevant members (active, deferred and pension credit members), with carve-outs for schemes mainly administered outside Great Britain and certain non-registrable schemes; smaller schemes may connect voluntarily under regulation 16 ^leg.

The Pensions Dashboards (Amendment) Regulations 2023 (SI 2023/858, in force 9 August 2023) made the single most important change: they revoked the original staging profile (Schedule 2), with its multiple staging deadlines and connection windows, and replaced it with one statutory connection deadline of 31 October 2026 for all relevant schemes. The amendment also reset the “reference date” to the scheme year end falling between 1 April 2023 and 31 March 2024, removed the early-connection process, and extended the deferral application window ^leg.

Liability for failure cannot be socialised. The Pensions Dashboards (Prohibition of Indemnification) Act 2023 (c.14) prohibits trustees and managers being reimbursed from scheme funds for civil penalties under section 238G ^leg. It was commenced on 1 January 2024, amending section 256 of the Pensions Act 2004 so that paying such a penalty from scheme funds is itself an offence, backed by the section 10 Pensions Act 1995 civil penalty regime ^leg.

Connection timetable and deferral

Although the legal backstop is a single date, DWP published a staged “connect by” timetable in August 2025 to manage the flow of connections, prioritising the largest schemes and providers so that the public dashboard can launch with high coverage as early as possible (Part 1 large schemes, then Part 2 medium schemes) ^DWP. The timetable is not mandatory, but trustees and managers must have regard to it under regulation 15(2)(d); failing to do so is itself a breach, and connecting too close to 31 October 2026 is flagged as a material risk of missing the legal deadline ^DWP.

Scheme size for this purpose is fixed at a historic reference date, not a current headcount: for occupational schemes it is the number of active, deferred and pension credit members at the scheme year end falling between 1 April 2023 and 31 March 2024, and for FCA-regulated providers the number of pots in accumulation at the scheme year end in the same window ^DWP. A scheme that has since shrunk below a band threshold therefore keeps its original cohort date.

The staged dates themselves are ^DWP:

Part 1 — large schemes and providers

Part 2 — medium schemes and providers

The Parliamentary pension scheme is staged separately, with its own dates of 31 January 2026 (750–999 members), 30 June 2026 (195–249) and 31 August 2026 (125–154) (same source). Schemes with fewer than 100 relevant members are outside the staged timetable and may connect voluntarily, and a scheme established on or after 1 April 2024 connects six months after the end of its first scheme year or on the date assigned to similar schemes, whichever is later (same source).

Two points follow for anyone reading this in October 2026. Every staged date has now passed — cohort 2(i), the smallest band in the timetable, was due by 30 September 2026 — so there is no longer any cohort still working to a guide date, only the statutory backstop of 31 October 2026. And the timetable’s own logic cuts against late connection: the guide dates were set to spread load across the ecosystem, so a scheme connecting in the final weeks is doing so in the window DWP designed the staging to avoid.

Schemes may apply to defer connection by up to 12 months (no later than 31 October 2027), but only on narrow grounds: that before 9 August 2023 they had embarked on a data-transfer programme to a new administrator and/or were contractually committed to retender administration, and that meeting the deadline would be disproportionately burdensome or put member data at risk. A change of administrator alone is insufficient, deferral can be granted only once, and the application deadline was 8 August 2024 ^DWP. A mere change of administrator does not require a formal deferral application provided connection completes by the deadline ^DWP.

The timetable that was revoked — and the staging logic that came back

The 2025 “connect by” timetable above is not an innovation; it is a return to the design the 2023 amendment had abandoned. DWP’s consultation on the draft Regulations (issued 31 January 2022, responses by 13 March 2022, published on GOV.UK under a 15 July 2022 page date) set out a three-wave staging plan: wave 1 large schemes of 1,000+ relevant members staging over April 2023 to September 2024, wave 2 medium schemes of 100 to 999 over October 2024 to October 2025, and wave 3 small and micro schemes of 99 or fewer, whose dates were “not to be set in these Regulations, but we expect this is likely to be from 2026” ^DWP. Draft Schedule 2 (regulation 14(2)(a)) carried 13 large-scheme cohorts and 11 medium-scheme cohorts, against a reference date of the scheme year end falling between 1 April 2020 and 31 March 2021 — three years earlier than the reference window the amendment later substituted:

Wave 2 then ran in eleven monthly steps: 850–999 by 31 October 2024, 750–849 by 30 November 2024, 600–749 by 31 January 2025, 500–599 by 28 February 2025, 400–499 by 31 March 2025, 320–399 by 30 April 2025, 250–319 by 31 May 2025, 195–249 by 31 July 2025, 155–194 by 31 August 2025, 125–154 by 30 September 2025 and 100–124 by 31 October 2025 (same source). The connection window as drafted was three months ending on the deadline for cohort 1(a) and one month for every other cohort (same source). (Citation caveat: the staging table’s scheme-type cells are column-flattened in the capture, so where a cohort lists two scheme types the row-by-row pairing of those two phrases is not recoverable from this file alone; the cohort letters, member bands and dates are unambiguous.)

Three things follow for anyone reading the current regime. First, the revoked profile would have finished a year earlier than the single deadline that replaced it — all relevant schemes connected by 31 October 2025, against the statutory 31 October 2026 now in force. The 2023 reset is usually described as a simplification, and it was; but on these dates it also bought roughly twelve months. Second, the 2025 guide timetable reproduces the same architecture — largest money purchase first, public service as a block, smallest occupational schemes last — on dates shifted about eighteen months to two years later. The staging logic was never thought wrong, only unachievable on the original dates. Third, slippage had begun before the Regulations were even made: the same GOV.UK page records post-consultation changes moving master trusts of 20,000+ from 30 June to 31 August 2023, money purchase AE schemes of 20,000+ from 31 July to 30 September 2023, public service schemes from 30 April to 30 September 2024, and the first cohort’s connection window from three months to five (same source).

The staging design rested on a concentration argument that is worth recording, because it is the clearest statement in the corpus of why order mattered more than speed. The roughly 1,000 large schemes in wave 1 were “just over 3% of UK schemes” but held 99% of active and deferred memberships; the roughly 2,000 medium schemes were about 6% of schemes and 1% of entitlements; and the over 27,000 small and micro schemes were about 90% of schemes but 0.26% of entitlements (same source). On that basis DWP expected nearly 99% of active and deferred memberships to be findable by the end of September 2024, 95% of money purchase memberships by end July 2023 and 95% of non-money purchase by end August 2024 (same source). Those coverage dates are now two to three years past, and the connection work was still being completed in 2026 — so the concentration logic held but the calendar did not.

Two drafting details carried forward differently from the draft. On hybrid schemes, draft regulation 15 set the deadline as the earliest of the deadlines for the scheme’s money purchase and non-money purchase sections taken separately (the consultation’s worked example: 9,000 non-money purchase plus 1,000 DC-for-AE members staged at February 2024 rather than June 2024, four months earlier), whereas the post-consultation summary on the same page moved to assessing the total relevant members across both sections and treating the whole scheme as non-money purchase (same source). And draft Schedule 3 Part 2 contained a small-pot exemption from projected value data where accrued money purchase rights were “less than £5,000” at the last illustration date, no contributions had been made since and the member had been notified — with projections also switched off within two years of normal pension age (same source); this is the draft, not the law as made, and should be checked against SI 2022/1220 before reliance.

(Corpus gap: no captured source records what became of wave 3. The 2022 consultation expected small and micro schemes of 99 or fewer relevant members to be staged “likely to be from 2026” in a later instrument, and schemes under 100 members remain outside the staged timetable and voluntary-only today — but nothing in raw/ holds a DWP decision, consultation or statement either setting wave 3 dates or abandoning the plan. Since that is roughly 90% of UK schemes, it is the largest open question about the ecosystem’s eventual coverage.)

Data readiness — TPR’s central message

TPR’s consistent theme is that member data is the binding constraint, not the plumbing. Connection requires schemes to match dashboard users to their records and return recent, accurate value data within strict timescales — values must derive from a statement or calculation within the last 12–13 months, returned within three working days for DC and ten working days in all other cases ^legTPR.

Through successive regulatory initiatives — low scheme-return scores (2024), the largest schemes representing ~80% of in-scope memberships (2025), and a 2026 initiative targeting 240 private-sector DB and hybrid schemes — TPR has found matching (“find”) data progressing well but value data lagging, especially for DB and hybrid schemes that have no annual benefit statement obligation and so hold staler values ^TPR. In November 2025 TPR consolidated its member-data guidance and told trustees to treat data as their “most important strategic asset,” warning of an industry-wide “data debt” — one in four schemes still hold non-digitised records and fewer than three in five are confident in their common data ^TPR. Trustees remain accountable for data quality even where tasks are delegated to administrators, and schemes unable to demonstrate compliance risk improvement notices and other formal intervention ^TPR.

By May 2026, about 75% of member records were already connected ^TPR, and by July 2026 around 85% were connected, with schemes and providers “on course” to meet the 31 October deadline ^HoC. Insights from the DB/hybrid initiative will help inform the timing of the MoneyHelper dashboard’s public launch (the “Dashboards Available Point” under regulation 4, which requires at least six months’ notice) ^legTPR.

TPR’s evidence sharpens the readiness picture. Its Market Oversight: Pensions Dashboards report (April 2026) reviewed the 51 largest schemes (about 80% of records), finding matching data far more mature than value data — few schemes had even received a readiness report from their integrated service provider — which is why TPR launched the follow-on DB/hybrid value-data phase and tied the findings to the dashboard launch decision ^TPR. A separate readiness-tracking survey put numbers on the gap: 76% of schemes were confident of connecting on time and 94% were satisfied with their matching data, but 36% of DB schemes held value data not calculated in the last 12–13 months, 35% gave members no online access to values, and member-portal registration typically ran at only 25–49% ^TPR. See TPR Market Oversight and prudential supervision.

Launch sequencing — and who regulates the dashboard itself

Connection is a deadline; launch is a decision, and the two are deliberately decoupled. Trustees planning member communications need the second timetable as well as the first, and the House of Commons Library’s September 2026 briefing sets it out more plainly than any scheme-facing source in the corpus ^HoC.

The enabling legislation — the Pension Schemes Act 2021 — allows for both public and private sector dashboards. In October 2024 the government said it would prioritise the MoneyHelper Pensions Dashboard, developed by the Money and Pensions Service (an arm’s-length body of DWP), before turning to private-sector dashboards. It still intends to enable private organisations to build and operate their own, and the Pensions Dashboards Programme is setting up a group of prospective operators to work on that development; all dashboards, public and private, will connect to the single digital ecosystem the PDP is building ^HoC.

So the sequence trustees should plan against is:

• 31 October 2026 — the statutory connection deadline ^leg. MaPS will give “a further update on launch plans around the connection deadline at the end of October” ^HoC.
• Six months’ notice — industry will receive six months’ notice of the launch date, mirroring the statutory requirement that the Secretary of State give six months’ notice of the dashboards available point ^HoCleg.
• 2027/28 — the window in which MaPS expects the MoneyHelper Pensions Dashboard to be available; private-sector dashboards are expected to follow ^HoC.

That gap matters operationally. A scheme connects in 2026 but faces no member-driven find/view volume until the dashboards available point, which on the current expectation is at least a year later — and it gets six months’ warning of it. The planning implication runs the other way from most compliance deadlines: the connection project must finish first, but the member-enquiry capacity it creates (the possible-match resolution route in particular) has to be live and resourced from the launch date, not from connection day.

Regulatory responsibility is split, and the dashboard itself is FCA territory. TPR and the FCA regulate the pension schemes and providers sharing data with dashboards; the FCA separately authorises and regulates pensions dashboard operators ^HoC. For a trustee board that means the scheme-side duties below are TPR’s (and, on data quality, the ICO’s), while the conduct of whatever front-end a member actually uses — including any commercial dashboard — is supervised by the FCA under its own authorisation regime, not by TPR. See Who regulates UK pensions — the bodies and how they fit together and FCA Consumer Duty and pensions.

How connection actually works — the operational mechanics

Behind the connection deadline sits a precise machine defined by Part 3 of the 2022 Regulations. When the Money and Pensions Service sends a scheme a find request, the trustees must run matching immediately against criteria the scheme itself decides and records (kept for six years) ^leg. The outcome is one of two states:

• A “match made” — the data meets the criteria well enough to confirm membership. The scheme registers a pension identifier (PeI) with MaPS, and on a later authorised view request (after MaPS confirms the individual’s consent) returns the member’s view data ^leg.
• A “possible match” — only a partial match. The scheme returns a limited message (its name and a contact) telling the individual to get in touch within 30 days; if they don’t, or the match can’t be resolved, the scheme must delete the find data and de-register the PeI ^leg.

View data comes in three categories: administrative data (scheme name, benefit type, membership status, administrator contact) returned immediately; signpost data (links to the SIP, implementation statement and any DC costs/charges); and value data (the projected/accrued figures, with contextual information). Value data must be current — from a statement within the last 13 months or a calculation within the last 12 — and is returned immediately if already current, otherwise within three working days for DC or ten working days for non-DC benefits ^leg. The matching task is, in GAD’s analogy, an elimination exercise against the scheme’s records using verified identity attributes (name, date of birth, address) and self-asserted data (National Insurance number, previous addresses), where “data quality issues, ranging from certainty to completeness, are commonplace” ^GAD.

Matching in practice — TPR’s August 2026 guidance

The regulations say a scheme must match; TPR’s “Matching people with their pensions” guidance (26 August 2026) says how, and it is the first TPR material in the corpus to treat matching as a governance discipline rather than a data task ^TPR.

The find request now has a known shape. Users are verified through GOV.UK One Login before they can issue one, so schemes “can be confident the user is who they say they are”. The core identity information that arrives is first name, last name, date of birth, email address and — only where it was used for two-factor authentication, so it may simply be absent if the user chose an authenticator app — a mobile number. GOV.UK One Login separately checks that any UK address supplied “exists and has an association with the user, through credit records”. Beyond that, the only additional user-supplied item is currently the National Insurance number, with previous or alternative surnames and further contact details flagged as possible future scope ^TPR.

The scheme sets its own matching criteria, calibrated to its own data. TPR anticipates “many schemes will use last name, date of birth and National Insurance numbers”, and notes criteria can be widened — first name, alternative names, personal email, postcode — to raise confidence ^TPR. Two duties attach to that freedom, and both are the kind trustees discover late:

• A six-year record. “You must keep a record of your matching policy for at least six years from the end of the scheme year in which the decision is taken.” Only the scheme-level criteria need recording, not the criteria applied to each individual search — but the policy itself must cover how possible matches and queries are managed and resolved, how often provider matching-performance reports are reviewed, the governance process monitoring whether the criteria work, and how and how often they are reviewed ^TPR.
• A DPIA is mandatory, not advisable. Matching, combining or comparing data from multiple sources “requires a Data Protection Impact Assessment (DPIA) under UK GDPR” — schemes must produce one or update an existing assessment (same source). See Data protection and cyber resilience for pension schemes.

TPR is also explicit about who enforces what. Poor data risks “returning data for the wrong person or not finding a pension record when you should”, which “may lead to enforcement action being taken against you by us or the Information Commissioner’s Office (ICO)” ^TPR. Dashboards data quality is therefore a two-regulator exposure, not a TPR-only one — a point the penalty figures below understate. Where data-quality problems surface, schemes “need to take targeted action to resolve them and document this in a data quality improvement plan”, with proactive fixes suggested: scans for missing or temporary NI numbers, employer outreach, tracing exercises, and collecting email addresses and mobile numbers (same source). See Record-keeping (TPR General Code module).

On outcomes, the guidance hardens the mechanics the regulations sketch. A match made requires the scheme to create and register a Pension Identifier (PeI) with the digital architecture — it carries no pension data, only confirmation that a match exists — and to de-register it “as soon as possible” when a member leaves the scheme or retires ^TPR. For a possible match (TPR’s example: NI number and date of birth match but the surname does not, because of an undisclosed change of name on marriage) the scheme registers a PeI with “possible” status and messages the user. The user then has 30 days to make contact; if they do, the scheme “must resolve the possible match promptly, including during user testing”, and if they do not — or the scheme “is unable to resolve the match in a timely manner following contact” — the scheme must delete their personal information and deregister the PeI (same source). Where no match is found at all, deletion is again mandatory: there are “no legal reasons to keep it and you will otherwise breach the GDPR requirements” (same source).

Two operational consequences follow that trustees should plan for rather than discover. First, resolving possible matches needs a staffed route and an identity check — TPR expects the scheme to decide how it verifies anyone responding to a possible-match query, and offers an enquiry form, a dedicated query-resolution team, or “an AI-enabled web-based solution” as the options ^TPR. Since the 30-day clock and the “promptly” standard run against the scheme, this is a resourcing commitment with a deadline attached. Second, multi-section schemes cannot assume one answer: where sections differ in data quality a scheme may need more than one set of matching criteria, and must decide in advance what to tell a member matched in one section but only possible-matched in another, and which contact details to give where different administrators serve different sections. Complex benefit structures may warrant a data-mapping exercise and, “where necessary”, updated contractual agreements with providers (same source) — which is the point at which this becomes an ISP contracting question as much as a data one.

Enforcement has teeth. TPR may issue compliance notices (and third-party compliance notices to an ISP or administrator whose act or omission caused the breach), followed by penalty notices of up to £5,000 for an individual and £50,000 for a body corporate — and, where there is consent, connivance or neglect, a penalty can be issued to a director or manager personally ^leg. Trustees must use “every endeavour” to stay connected and notify MaPS of downtime or cyber incidents ^leg.

TPR’s before-and-after-connection checklists translate this into trustee actions: before connecting — put effective oversight in place, confirm the connection deadline, choose between a direct build and an integrated service provider (ISP) route, sort supplier contracts, and run a data-quality programme covering both matching data and value data (including a Data Protection Impact Assessment); after connecting — stay connected, embed data-quality controls, maintain board oversight and meet reporting and record-keeping duties ^TPR.

TPR published these as a dedicated checklist page in September 2026, and the detail there is worth reading against a project plan built from the overview. Three things sit differently from where trustees usually put them. Reporting and record-keeping is a before-connection action — understand what must be reported and what records must be kept, and put the processes in place, rather than standing them up after go-live. The delivery-plan conversation is named party by party — software provider, actuary, legal adviser, employer and AVC provider — with regular implementation updates expected; AVCs are the benefit most often missing from a scheme’s data plan. And the after-connection reporting duty does not end with the provider filing: trustees are to ensure providers submit reporting data to MaPS when required, obtain a copy of those reports, and keep records available to TPR on request. The same page frames staying connected as an ongoing requirement whose lapse breaches the law, asks trustees to confirm that all relevant members are connected, and expects the processes — matching in particular — to be reviewed and refined in light of testing ^TPR. The full item-by-item version is in Pensions dashboards connection checklist.

A corpus caveat for practitioners. The regulations and DWP’s staged timetable (set out above) are necessary but not sufficient: the binding technical specification — the PDP Code of Connection and the data, technical and design standards — is published by MaPS/PDP and approved by the Secretary of State under regulation 5 ^leg. Those standards sit on the Pensions Dashboards Programme site, which this corpus does not hold (the site’s terms prohibit text and data mining), so a scheme cannot achieve operational compliance from the regulations alone — read them alongside the current PDP standards on that site.

Related

Dashboards interact with the wider reform agenda — see Automatic Enrolment (which expanded the pool of small, deferred pots dashboards aim to surface), The Pensions Investment Review — scale, consolidation and Value for Money, The Pension Schemes Act 2026 and DB surplus / endgame policy, Collective Defined Contribution (CDC) / Collective Money Purchase Schemes, and LGPS Reform — Pooling, Governance, Access & Fairness, and Valuations.