Why a client portfolio is a different problem
The calculation is not the hard part. The hard part is that the inputs arrive incomplete, undocumented, and with a method nobody chose deliberately.
Someone managing their own portfolio can go back and check their history. A client walking in with a folder of screenshots from a platform that no longer exists has nothing to go back to. Most of the work in an advisory file is establishing what is actually knowable and what is now guesswork.
These notes are written from that angle. They apply whatever the asset: a bitcoin tax file is the same exercise as one for any other token, because the mechanics are identical. What varies is the venue, the record quality and the tax year. They cover the questions that come up repeatedly with digital asset clients, and the ones worth resolving before you produce a number.
Lot-level tracking, and where it stops
Most venues report aggregated transactions, which is enough to reconstruct lots yourself if the data is complete. It is not the same as receiving a lot-level history directly, and the distinction matters for two reasons.
First, specific identification requires that every disposal can be traced to a specific acquisition. If the venue collapsed trades into a single daily total, that trace does not exist and cannot be invented afterwards.
Second, a client who claims specific identification used has to produce something that supports it. On a typical venue export, the support is the reconstruction itself, which has to be done consistently and documented. That reconstruction is the same work FIFO would require, done for a method that gives less room for error.
A reasonable position with a client is to establish early whether lot-level history exists. If it does not, the available methods narrow, and the election should reflect that rather than the other way round.
For a CPA: what the method election is really for
A CPAs first question is usually how much of the figure is defensible. For digital assets that comes down to three things: whether the client chose the method deliberately, whether the records support that choice, and whether anyone wrote down what was assumed at the time.
An accountant or adviser who cannot answer those three questions is relying on the reconstruction having been luck as much as on it having been done correctly.
Take a position on the method, and write it down
The single most common failure is a file where the method is not stated and cannot be inferred. Two years later nobody knows whether the basis figures were FIFO or specific identification, and the answer changes the result materially.
- Establish what is actually permitted Availability of LIFO or specific identification varies and some jurisdictions restrict switching. Confirm before applying anything.
- Check whether the data supports the method Specific identification needs lot-level history. If the venue never provided it, the method is unavailable regardless of what it would have produced.
- Record the election and the date A dated note stating the method, the reason, and who decided is worth more than any calculation and costs nothing.
- Keep it consistent across the period Changing between periods to reduce a figure is the kind of thing that turns a routine file into an enquiry.
Method configuration is part of the work, not a setting. It should appear in the file in the same way a depreciation schedule appears in a fixed asset file.
What to request from a client
| Request | Why it matters |
|---|---|
| Raw exports, not summaries | A summary cannot be audited and usually cannot be corrected. Screenshots cannot be reprocessed. |
| Exports from every venue, full history | Most clients have forgotten an old account. Ask for the full range, not the recent period. |
| Statement of self-custody wallets | Movements in and out of wallets appear in no venue record and are frequently the source of a gap. |
| Any staking or DeFi activity | Income the venues did not report. Often not disclosed by the client because it does not look like income. |
| The client's stated intent on transfers | Which movements were internal and which were sales. Ambiguous transfers are the biggest source of phantom gains. |
| Price source for income events | Needed because the value on the receipt date is otherwise unrecoverable. |
Asking for these in one message, before starting, costs less than reconstructing them afterwards. Most of what makes a digital asset file difficult was knowable at the outset.
Establishing what is knowable
Work forwards from the final holdings and backwards from the records. Compare the token quantities the client holds against what the transactions explain.
Where the two disagree, the difference is income nobody recorded. That is not a client failure of diligence so much as a failure of the platforms to report, and it is worth stating plainly in the file rather than absorbing it into a calculated figure.
Where a venue is unreachable, say so and document the attempt. A known gap recorded at the time is defensible in a way that a silent gap discovered after submission is not.
A reconciliation routine worth running every time
- Export counts reconciled against the venue's own transaction history, with date ranges confirmed
- Final balances reconciled against trades plus recorded income
- Withdrawals matched to deposits across venues, with every unmatched row investigated
- One token traced end to end by hand, lot by lot
- Roll boundaries excluded from any relative measure
- Fee treatment checked on both the acquisition and disposal side
Steps one to three find the problems. Step four finds the ones that survive. A full pass takes under an hour on a portfolio of any realistic size and removes the majority of corrections that otherwise surface after filing.
When the answer depends on the client
Several things turn on facts only the client holds, and these cannot be resolved from the records:
- Whether a wallet is the client's, or held for someone else
- Whether a movement was a sale to a third party or an internal transfer
- Whether a token was received with a genuine disposal right at that moment
- How long each holding was held, where the rules distinguish periods
- Whether a wrapping or migration involved a genuinely different asset
Record the client's answer and the date it was given. Those five items are the boundary between what can be calculated and what has to be taken on trust.
Limits of these notes
General reference material on how the calculation works. Not advice, not a substitute for jurisdiction-specific guidance, and not a review of any individual position.
Digital asset treatment is changing quickly in most jurisdictions, particularly around staking income and decentralised finance. Confirm current requirements against the applicable authority before relying on any of it, and take specialist advice where the position is genuinely novel.