Specialist US UK Tax Services: Rebuilding Lost Records

When a dual US-UK filer needs three years of US returns, six years of FBARs and several years of UK returns prepared, the binding constraint is almost never the tax technical work — it is evidence. Specialist US UK tax services exist largely to solve that problem: reconstructing lost statements, contract notes and payroll records to a standard that survives both IRS and HMRC scrutiny.

At Jungle Tax the single most common misconception we correct in a first meeting is the assumption that a multi-year catch-up is a return-preparation exercise. It is not. It is a forensic records exercise with a return-preparation step bolted on the end. The returns themselves — three Forms 1040 with their international schedules, six Reports of Foreign Bank and Financial Accounts, and however many Self Assessment years HMRC is entitled to reach — are the last four weeks of a six-month project. Everything before that is finding, rebuilding and defending numbers that no longer exist in documentary form.

Why record recovery, not the returns, sets the timetable

A high-net-worth cross-border position typically involves ten to thirty reportable accounts across two jurisdictions: current and savings accounts at UK clearers, an investment platform or two, a stocks and shares ISA, one or more legacy workplace pensions, a SIPP, an employer share plan, a US brokerage account left behind at emigration, possibly a US retirement account, and occasionally a UK limited company. Each of those sits with a different institution, each institution has a different archive depth, a different retrieval process and a different response time, and none of them are in a hurry.

You cannot compress that. A platform can restore ten years of consolidated tax vouchers in forty-eight hours; a bank that closed an account in 2017 may take eight to twelve weeks to answer a data subject access request, and may come back with transaction listings rather than statements. A pension administrator that has changed hands twice since the member left service may take a full quarter to confirm contribution history. The critical path through a multi-year catch-up runs through the slowest institution in the chain, and the professional skill lies in identifying that institution in week one and starting it moving before anything else happens.

This is also why sequencing matters. Requests go out in parallel, not in sequence, and they go out before any technical analysis begins. Deciding whether a fund is a passive foreign investment company is worthless if you cannot obtain the acquisition dates. The order of operations in a properly run catch-up is: inventory, request, reconstruct, analyse, prepare, certify.

What "missing records" actually means in a cross-border catch-up

The US side of the gap

For the US return, the gaps that matter are rarely income totals — those can usually be rebuilt. The gaps that hurt are: cost basis and acquisition dates for securities and funds; the date and sterling value of every purchase into a collective investment scheme, because that drives passive foreign investment company computations; the maximum balance in each foreign account in each calendar year, which is the sole datum the FBAR actually needs; foreign tax paid by year and by category, for foreign tax credit purposes; and pension contribution and growth history where a treaty position is being taken.

The UK side of the gap

For Self Assessment, the missing items are usually dividend and interest vouchers, contract notes supporting capital gains computations, rental income and allowable expenditure for a let property, employment records where a P60 or P11D has been lost, and evidence of amounts remitted where a remittance-basis position was previously claimed. Following the replacement of the remittance basis with the four-year foreign income and gains regime from 2025-26, older remittance-basis years still require the historic evidence even though the regime no longer applies prospectively — a point generalist guides consistently miss.

The evidence hierarchy: what the IRS and HMRC will actually accept

Not all reconstruction is equal. We work to a four-tier hierarchy, and every reconstructed figure in a filing pack is tagged with the tier it came from. This is what allows a reviewer — whether an IRS examiner picking up a streamlined submission or an HMRC officer reviewing a disclosure — to see immediately that the taxpayer took reasonable care rather than guessed.

Tier	Evidence type	Examples	Weight in a filing
Tier 1	Primary contemporaneous documents	Original statements, contract notes, consolidated tax vouchers, P60s, completion statements, custodian year-end packs	Definitive. Use wherever obtainable, at any cost.
Tier 2	Institutional reissue or authority-held data	Reissued statements, IRS wage and income transcripts, HMRC employment history, registrar dividend histories, Companies House filings	Near-definitive. Treated as primary by both authorities.
Tier 3	Corroborated derivation	Balances derived from adjacent-year statements plus transaction listings; basis derived from published fund prices at a documented acquisition date; income derived from bank credits matched to a payer	Acceptable when the derivation method is disclosed and reproducible.
Tier 4	Reasoned estimate	Best-judgement figures where no record survives, built from a documented assumption set and deliberately conservative	Acceptable only if flagged, explained, and consistently biased against the taxpayer.

The discipline that matters is never mixing tiers silently. A schedule showing a maximum account balance with no indication that it was derived rather than read off a statement is a credibility risk. The same figure, footnoted as "derived: highest quarter-end balance in transaction listing, uplifted to the highest single-day credit", is a strength.

Where the records actually are: US recovery routes

The IRS holds more about a US taxpayer than most people expect. The wage and income transcript reports information returns filed under a Social Security number — Forms W-2, 1099 in its various flavours, 1098 and 5498 — and generally reaches back around a decade. For an American who left the US mid-career, this frequently recovers the entire US-source side of the picture: the brokerage 1099-B and 1099-DIV, the residual employer W-2, the IRA custodian's 5498s. Transcripts can be pulled through an IRS online account or requested on Form 4506-T, which also delivers account transcripts and verification of non-filing — the latter being useful evidence that a year genuinely was not filed.

Beyond transcripts: US custodians typically retain statement archives online for several years and will restore beyond that on request; transfer agents hold dividend and reinvestment history for directly registered shares, which is often the only surviving source of basis for legacy holdings; and prior-year returns, even unsigned drafts recovered from a former preparer, establish elected positions and carryovers that would otherwise be lost. Recovering a prior foreign tax credit carryover schedule can be worth more than every other document in the file combined.

Where the records actually are: UK recovery routes

The UK recovery map is different in character. HMRC's own systems hold employment and PAYE history, which the taxpayer can view through their personal tax account and which we can obtain on their behalf under authority — this reliably rebuilds employment income, tax deducted and often National Insurance history for years where the P60 is long gone. Banks and building societies must retain customer records for a period after the relationship ends, and a data subject access request will usually produce transaction data even for closed accounts, though frequently as raw listings rather than formatted statements.

Investment platforms are the best-behaved source in the UK market: most will regenerate consolidated tax certificates and full transaction histories covering many years on a single phone call. Fund managers hold unit-level acquisition data. Share registrars hold dividend and corporate action history. Pension administrators hold contribution records, though legacy schemes that have been bought, merged or bulk-transferred are the single most common source of a hard, unrecoverable gap. Where a property is involved, the conveyancer's file and Land Registry records establish acquisition date and consideration even when the completion statement has vanished.

Reconstructing the genuinely hard items

How do you reconstruct an FBAR maximum account value with no statements?

The FBAR asks a narrow question: what was the maximum value of the account during the calendar year? It does not ask for income, and it does not ask for a year-end balance. That narrowness is helpful, because it means a full statement run is not strictly necessary — you need a defensible high-water mark.

The reconstruction ladder we use runs: actual statements first; then any periodic balance data the institution can produce, taking the highest observed point and uplifting for intra-period peaks evidenced in a transaction listing; then a roll-forward from a known balance at either end of the gap using the transaction history; then, where nothing survives, a conservative high estimate flagged as such. Because the FBAR is an information return rather than a tax computation, an over-stated maximum value carries no tax cost while an under-stated one is a compliance exposure. The bias is therefore always upward, and that asymmetry should be stated explicitly in the working papers. For a sense of what is at stake across multiple unfiled years, our FBAR penalty calculator models the exposure ranges.

Rebuilding cost basis when the contract notes are gone

Basis is where cross-border catch-ups get genuinely difficult, because a dual filer needs two bases for the same asset. HMRC computes the gain in sterling using the sterling cost at acquisition and the sterling proceeds at disposal. The IRS computes the gain in dollars, translating the acquisition consideration at the exchange rate on the acquisition date and the proceeds at the rate on the disposal date — which manufactures foreign-currency gain or loss on an asset that may have moved very little in local terms. Reconstructing one basis and assuming the other follows is the classic generalist error, and it produces returns that cannot be reconciled to each other.

Where contract notes are lost, basis can usually be rebuilt from platform transaction histories, registrar records, published closing prices at a documented trade date, corporate action histories for demergers and share splits, and employer share plan records for options and restricted stock. Where the acquisition date itself is unknown, the conservative approach — assuming the earliest plausible acquisition and therefore the lowest plausible basis — is both defensible and, importantly, adverse to the taxpayer, which is exactly the posture a reviewer wants to see. Our cross-border tax team maintains standing basis-reconstruction protocols for the UK platforms and registrars we encounter most.

ISAs, unit trusts and the passive foreign investment company problem

A stocks and shares ISA is invisible to HMRC and highly visible to the IRS. It is not a tax-free wrapper for US purposes; it is a taxable account holding what are almost certainly passive foreign investment companies, and each fund holding potentially requires its own Form 8621. The reconstruction burden here is unusually heavy, because the default excess distribution regime needs the full holding period: acquisition date, acquisition cost, every distribution in every year, and every disposal.

Two practical points that competitor guides miss. First, the elections that would make the position cheaper — qualified electing fund treatment, or mark-to-market where the fund is regularly traded — depend on information the fund manager may simply never have produced for a US shareholder, so the reconstruction has to establish not just what happened but what data exists. Second, where the holding pre-dates the first year of a streamlined submission, the prior-year data still has to be rebuilt, because the excess distribution allocation reaches back across the entire holding period regardless of which years are being filed. A three-year catch-up can require a decade or more of fund data.

UK pensions, payroll and employer records

Pension reconstruction serves two masters. HMRC needs contribution history to test annual allowance and any tapering; the IRS needs enough to support the treaty position taken on growth and, where distributions have begun, the character and source of those payments. Where a legacy scheme has been transferred, the receiving administrator generally holds the transfer value but not the underlying contribution history — a genuine dead end that must be handled with a documented estimate rather than a silent omission.

Payroll is easier than clients fear. Between HMRC's employment records, the employer's own retention obligations and, for the US side, the wage and income transcript, employment income is one of the few categories that can almost always be rebuilt to Tier 1 or Tier 2 quality. Benefits in kind reported on a P11D are the weak point, and where the form is lost the employer's payroll provider is usually the fastest route.

US versus UK: how the two authorities treat reconstructed figures

Issue	United States / IRS	United Kingdom / HMRC
Formal position on estimates	No general statutory estimate rule, but reasonable reconstruction supported by third-party data is long-established practice; strict substantiation applies to certain categories of deduction	Published guidance expressly contemplates estimation where records are incomplete, provided the taxpayer estimates as accurately as possible and retains the workings
Disclosure of the estimate	Best practice is a statement of method within the submission; unflagged estimates undermine a non-willfulness certification	Estimated figures should be identified on the return or disclosure, with the basis explained
Years typically in scope	Three years of income tax returns and six years of FBARs under the streamlined procedures	Driven by taxpayer behaviour, with materially longer assessment windows available for offshore matters and for failure to notify
What the reconstruction is judged against	Whether the failure was non-willful, certified in the relevant Form 14653 or 14654 narrative	Whether reasonable care was taken, which drives the penalty behaviour category and the disclosure period
Consequence of a weak reconstruction	Submission may be treated as incomplete and the protection of the programme can be lost	Behaviour may be recharacterised upward, extending the years in scope and increasing penalties
Currency of the underlying record	US dollars, with date-specific or published average translation depending on the item	Sterling throughout, with no equivalent translation layer

The strategic point is that both authorities are assessing the same thing through different vocabulary: did this taxpayer make a genuine, methodical, good-faith effort to get it right? A well-documented reconstruction is therefore not merely an accounting convenience. It is the primary evidence for the behavioural conclusion on which both US programme eligibility and the UK penalty position turn.

The reconstruction file: the deliverable nobody talks about

Alongside the returns, a properly run catch-up produces a reconstruction file that is never filed but is retained and can be produced on twenty-four hours' notice. It contains, for every account and every year: the source of each figure, the tier it sits at, the derivation where the tier is 3 or 4, copies of every request made to every institution with dates, and copies of every refusal or partial response received.

Those refusals matter more than clients expect. A letter from a bank confirming that records before a given date no longer exist is the single most valuable document supporting a Tier 4 estimate, because it converts "the taxpayer guessed" into "the taxpayer exhausted the available evidence and then estimated". We routinely request written confirmation of archive limits even when we already know the answer, precisely so that the file contains it.

How does reconstruction support the non-willfulness narrative?

The streamlined procedures require a certification that the failure to report was non-willful — the result of negligence, inadvertence, mistake, or a good faith misunderstanding of the law. The IRS guidance on the streamlined filing compliance procedures sets out that framework, and the certification narrative is where a reconstruction either helps or hurts.

A narrative that says "I did not realise my ISA was reportable" is thin. A narrative that says the same thing and is accompanied by a submission demonstrating that the taxpayer approached eleven institutions, recovered what could be recovered, obtained written confirmation of what could not, and estimated conservatively where necessary, tells a coherent story about a person who was unaware rather than evasive. The reconstruction is the corroboration for the assertion. Guides that treat the narrative and the numbers as separate workstreams have this backwards. Our streamlined filing specialists draft the two together, from the same working papers.

The UK mirror image applies where a Worldwide Disclosure Facility route is used: notification starts a defined window in which the full disclosure must be delivered, and that window is short relative to institutional response times. Notifying before record recovery is well advanced is one of the more expensive sequencing errors available, because it can force reliance on Tier 4 estimates that better planning would have avoided.

Sequencing the US and UK sides so the credits work

There is a further reason record recovery drives the timetable: the two returns are not independent. Foreign tax credit claims on the US return depend on UK tax actually paid and on the year in which it is treated as paid or accrued. If the UK years are being amended or disclosed at the same time as the US years are being filed, the UK liability is a moving number, and a US return prepared against a UK figure that later changes will need amending — adding a further cycle to a project that is already long.

In practice we fix the UK position first wherever the facts allow, then build the US returns against a settled UK liability, then revisit only if HMRC disturbs the disclosure. Where the UK position genuinely cannot be settled first, we prepare the US returns on a documented provisional basis and plan the amendment cycle into the engagement from the outset rather than discovering it later. This is standard practice in the high-net-worth cross-border context and almost entirely absent from generalist expatriate guidance.

A realistic timetable

• Weeks 1-2 — inventory and request. Build the complete account and entity inventory from memory, credit files, correspondence and any surviving paperwork. Issue every institutional request simultaneously. Lodge transcript requests. Obtain HMRC authority and pull employment history.
• Weeks 3-8 — collection and chasing. Institutions respond unevenly. Escalate the slow ones early. Identify which gaps are permanent and start collecting written confirmations of archive limits.
• Weeks 6-12 — reconstruction. Rebuild balances, basis and income by account and by year, tagged by evidence tier. Resolve fund-level data for any PFIC analysis. Settle the exchange rate methodology and apply it consistently across every year.
• Weeks 10-16 — analysis and preparation. UK computations first where possible, then the US returns, international schedules, FBARs and Forms 8621 and 8938.
• Weeks 14-20 — certification and submission. Draft the certification narrative from the reconstruction file, review, sign, file.

Twenty weeks is a well-run case with cooperative institutions. A case with a merged pension scheme, a closed offshore bank and a fifteen-year fund holding runs longer. What compresses the timetable is not working faster at the end; it is issuing better requests at the start.

The mistakes that cost the most

• Filing what you can prove and staying silent on the rest. An account omitted because the records were unavailable is an omission, not a gap. It is visible under automatic exchange of information and it undermines the entire submission.
• Starting a formal disclosure clock before the records are in. Both the US and UK routes reward preparation and punish improvisation.
• Estimating in the taxpayer's favour. Every Tier 4 figure should be conservative. The cost of conservatism is small; the cost of an aggressive estimate discovered later is the loss of the behavioural argument.
• Using one basis for both returns. Sterling basis for HMRC, dollar basis with date-specific translation for the IRS. They are different numbers and both must be built.
• Discarding the workings once the returns are filed. The reconstruction file is the defence. It should outlive the returns by years.

Further reading across our cross-border guides covers the individual forms and disclosure routes in detail, alongside our US tax services and UK tax services for dual filers. General background on Self Assessment obligations and record-keeping is available from HMRC's Self Assessment guidance.

Speak to us before you start requesting records

The order in which requests go out, the wording used, and the decision about when to notify an authority all shape the outcome of a multi-year catch-up far more than the return preparation that follows. If you are facing three years of US returns, six years of FBARs and an unresolved UK position with incomplete records, the most valuable hour you will spend is the one at the beginning. Contact our cross-border team for a confidential consultation. We will map the evidence you hold, the evidence you can still recover, and the reconstruction methods that will carry the rest — before any clock starts running.