Stablecoin Reserve Reconciliation

The defining financial control of a stablecoin business is a reconciliation: tokens in issue on one side, backing assets held on the other, agreed daily, evidenced permanently. Everything else in the finance function — the accounting policies, the treasury controls, the audit, the coming regulatory returns — hangs off that one discipline, and it is the control every serious counterparty now asks about: auditors, banking partners, investors and, as the UK regime crystallises, the FCA. This guide sets out how to build the reconciliation to the standard the scrutiny will demand, drawing directly on the client-money discipline that CASS firms have run for years — because that is precisely the standard the stablecoin rules are converging on.

The principle: same discipline, new ledgers

A CASS firm reconciles its internal records of client entitlements to the client bank accounts holding the money. A stablecoin issuer reconciles the token ledger — what is in circulation and therefore redeemable — to the reserve portfolio backing it. The mechanics differ; the logic is identical: an independent, evidenced, frequent agreement between the obligation and the asset, with any difference treated as an incident rather than a rounding note. Issuers who frame the control this way inherit two decades of hard-won client-money practice — segregation, designation, daily cadence, breaks discipline — instead of inventing standards from scratch. Those who frame it as a treasury spreadsheet inherit the enforcement history that taught payments firms the safeguarding lesson the hard way.

The two sides of the reconciliation

The obligation side — tokens in issue. The source is the on-chain record: total supply, adjusted for tokens the issuer itself holds (minted-but-unissued, redeemed-and-unburned, treasury balances), across every chain the token is deployed on. Multi-chain issuance multiplies the reconciliation points, and bridge or wrapped arrangements need explicit treatment — decide, and document, whose supply counts. The obligation figure should be system-generated, not manually assembled: a query against nodes or a reliable indexer, snapshotted at a defined daily timestamp, with the extraction itself logged. The asset side — the reserve. The sources are custodian and bank statements: cash at credit institutions, short-dated government securities, money-market instruments — whatever the reserve policy permits — valued on a documented basis, at the same timestamp, with accrued income treated consistently. Independence matters on both sides: the reconciler should draw obligation data and asset data from sources the transaction-processing team cannot amend, which is the segregation-of-duties principle every controls framework starts from.

Cadence, tolerance and the breaks discipline

Daily is the defensible cadence — not because a rule yet says so in terms, but because redemption risk is daily and the client-money precedent the regime is converging on is daily. The reconciliation should run to a fixed timetable (say, 9am against the prior midnight snapshot), complete within a defined window, and produce a standard output: obligation, assets, surplus or deficit, and an itemised list of reconciling items. The breaks discipline is where reconciliations earn their keep. Every difference gets classified — timing (a redemption settled on-chain but not yet in the bank), valuation (price source divergence), or true break (neither) — with an owner, an ageing clock and an escalation threshold. Timing items that persist past their natural settlement window are re-classified as true breaks, not left to age politely. A deficit against the obligation is an incident with a same-day escalation path to whoever holds senior responsibility — defined in advance, in writing, because discovering the escalation route during the first deficit is how small problems become reportable ones. The parallel with CASS breaks handling is deliberate and complete.

Evidence: build the file as you go

Every reviewer of this control — auditor, regulator, diligence team — will ask the same question: prove it ran, every day, all year. The evidence pack that answers it: the daily reconciliation output, archived immutably with its underlying extracts; the breaks log with resolution trails; the documented methodology (sources, timestamps, valuation bases, tolerance and escalation thresholds, the multi-chain and bridge treatments decided above); attestations or sign-offs showing four-eyes review; and the change log when methodology evolves. Build this file daily and the annual audit reads it in a week; reconstruct it annually and the audit becomes archaeology — the identical lesson our guide to the CASS audit draws for client money, and the reason experienced practitioners treat evidence as part of the control, not a by-product of it.

Beyond the daily rec: the surrounding controls

The reconciliation sits inside a small family of controls that reviewers expect alongside it: reserve policy compliance monitoring (asset eligibility, concentration and diversification limits, maturity profile) checked on the same cadence; independent periodic attestation of the reserve, increasingly a market expectation for issuers regardless of regulation; treasury movement controls — dual authorisation on reserve transfers, whitelisted destinations, key-management arrangements finance understands and can evidence; and the month-end bridge, where the daily reconciliation ties to the general ledger so the statutory accounts and the reserve reporting tell one story — the discipline our guide to the regulated close describes. Together these are what “audit-ready” and “authorisation-ready” actually mean in practice.

Who builds and runs this

The build is a defined project — typically eight to twelve weeks for an experienced hand to design the architecture, stand up the data feeds, document the methodology and run it to stability — and the running is a daily discipline that belongs in finance, with compliance owning framework and challenge. The natural builders are CASS and safeguarding-experienced accountants, for whom this is a familiar control on unfamiliar ledgers, and regulated-audit alumni who have tested the equivalent from outside; pure crypto-native finance experience is welcome but scarcer than the crossover profile. Our crypto and digital assets practice recruits both, permanent and interim — and for pre-launch issuers, the interim build engagement ahead of go-live is usually the right first hire, exactly the shape of the mandates currently on our jobs board.

The failure modes we see

The same weaknesses recur across early-stage issuers, and naming them is the cheapest prevention. The spreadsheet single point of failure: one analyst, one workbook, no independent data feeds — fine until the analyst is on holiday during a depeg scare. Timestamp drift: obligation snapshotted at midnight, assets at whenever the custodian statement arrived, differences waved through as “timing” — the reconciliation that always balances because it never really compares. Treasury tokens counted ambiguously: minted-but-unissued supply drifting in and out of the obligation figure depending on who ran the query. Breaks without owners: a reconciling-items tab that grows monthly because nothing ages into escalation. Evidence assembled retrospectively: the control ran, mostly, but the file proving it gets built the week before the audit. Each of these is an architecture decision made lazily at the start; all are cheap to fix in the build and expensive to fix under scrutiny — which is the strongest argument for the experienced builder over the willing generalist.

The regulatory horizon, held lightly

This guide deliberately avoids citing rule numbers from a regime still moving through consultation, but the direction is settled enough to build against: backing requirements for issued tokens, segregation and custody expectations, redemption rights honoured at par and at pace, and reporting that evidences all of it — the client-money pattern, applied to tokens. Firms that build the daily reconciliation, the reserve policy monitoring and the evidence file described above will find the final rules largely descriptive of what they already do; that was precisely the experience of well-prepared payments firms under the safeguarding reforms, and it is the standard our regulatory reporting practice recruits toward. Build to the discipline, not the draft, and the rulebook holds no surprises.

A build sketch: the first thirty days

For teams starting from zero, the opening month of a competent build runs roughly: week one — map the flows (chains, custodians, banks, the redemption path) and decide the methodology questions above in writing; week two — stand up the data feeds and run the reconciliation manually, daily, however rough; week three — document, add the breaks log and escalation thresholds, introduce four-eyes review; week four — automate the extracts, archive the evidence immutably, and present the first month’s file to whoever holds senior responsibility as a dress rehearsal for the audit. Manual-first matters: automation built before the process is understood automates the misunderstandings. By day thirty the control exists; the following two months harden it to the standard the scrutiny will demand.