How to Reconcile Exchange and Wallet Positions Against Your Fund Administrator
Reconciling exchange and wallet positions against your fund administrator means comparing, per asset and per venue, what your ledger expects you to hold against what each exchange, wallet, and broker actually reports — then investigating every discrepancy above a set dollar threshold before it reaches an LP report. Below is the step-by-step process.
Why reconcile at the venue level instead of just the fund total?
Comparing only the fund's total NAV to the fund administrator's total NAV can hide a lot: a $30,000 shortfall on one exchange can be masked by a $30,000 overstatement on a wallet, netting to zero at the fund level while both underlying numbers are wrong. Reconciling per asset and per venue is what surfaces the individual breaks instead of a false all-clear.
The reconciliation process, step by step
- Pull live balances from every connected venue
Query every exchange, wallet, and broker connection for its current balance, per asset. This should run read-only against each venue's API or public chain data — never against a manually maintained list, which is always at risk of drifting from what venues actually hold.
- Compute the ledger-expected balance for the same date
Reconstruct what the double-entry ledger says each asset, on each venue, should hold as of the same valuation date — not from a live dashboard snapshot, but from the append-only journal itself, so the comparison is apples-to-apples.
- Diff live vs. expected, per asset and per venue
Compute the exact dollar difference for every (asset, venue) pair. A pure diff, not a judgment call — this is where a "the numbers feel off" complaint becomes a specific figure: "$1,842 short on ETH at Coinbase" rather than a vague sense that something's wrong.
- Grade each difference against a materiality threshold
Not every difference matters. A tolerance — often expressed relative to the asset's decimal precision, or a flat dollar materiality floor — separates rounding dust from a genuine discrepancy that needs a human to look at it. Breaks below the threshold can be auto-resolved; breaks above it need investigation.
- Investigate and resolve each material break
For each flagged break: check whether it's a timing difference (a transfer in transit), a pricing difference (one side using a stale or different price source), a sync failure on one venue, or a genuine error on either the manager's or the administrator's side. Document the resolution — resolved, ignored with a note, or corrected with a reversing entry — so there's an audit trail.
- Compare the reconciled shadow NAV against the administrator's official NAV
Once individual venue-level breaks are cleared, compare the fund-level shadow NAV to the administrator's published NAV. They should tie within a tight tolerance — if they don't, the discrepancy is now isolated to something that survived venue-level reconciliation, which narrows the investigation considerably. See why doing this daily beats waiting for the monthly close.
How often should this process run?
Daily, ideally, or continuously if the underlying connections support it. A process run only at month-end inherits every weakness of a monthly cycle — see the real failure modes that a slow reconciliation cadence misses. Running it daily means a sync failure or a stale price gets caught within a day, not discovered a month later during the administrator's close.
Nyx Fund automates this entire process across roughly 95 exchanges and 21 blockchains: it pulls live balances, reconstructs ledger-expected positions, computes typed breaks with materiality grading, and compares against your administrator — daily, not monthly. Run the free NAV Validator to see how a live diff would score your current process, and grab a sample AI-drafted LP report to see the output on the other side of a clean reconciliation.
Key takeaways
Reconcile per asset and per venue, not just at the fund-total level — netting errors can hide two large offsetting discrepancies.
Pull live balances directly from each venue's API or on-chain data, never from a manually maintained list.
Grade every difference against a materiality threshold so rounding dust doesn't drown out genuine breaks needing investigation.
Document every resolution — resolved, ignored, or corrected — to keep an audit trail an auditor or LP can review.
Running this daily rather than monthly is what actually catches a sync failure or stale price before it compounds.
Questions, answered
What's the difference between reconciling at the fund level vs. the venue level?
Fund-level reconciliation only compares total NAV figures, which can hide two offsetting errors that net to zero. Venue-level (and asset-level) reconciliation compares expected vs. actual balances for each specific exchange, wallet, and asset, surfacing every individual discrepancy rather than a misleading all-clear.
How do I decide what dollar threshold counts as a material break?
Most funds set a materiality floor — a flat dollar amount, such as $1,000, or a small percentage of NAV — below which differences are treated as immaterial rounding and above which they require investigation. The right threshold depends on fund size and LP expectations, but it should be documented and applied consistently rather than judged case by case.
What should I do when a reconciliation break can't be explained?
Do not resolve or ignore it without a clear cause — an unexplained break should stay open, escalate to the administrator if it involves their figures, and be documented until a cause is found. Silently writing off an unexplained discrepancy is one of the most common ways real problems go undetected.
Can this reconciliation process be automated for a small fund?
Yes. The steps — pulling live balances, computing ledger-expected positions, diffing, and grading materiality — are mechanical and well suited to automation. Nyx Fund runs this process for emerging crypto and multi-asset funds without requiring a dedicated operations team.