How to review an incomplete tax file: not found, single source and mismatched information

Published 23 August 2026. Last reviewed 23 August 2026. Technically reviewed by the Taxpartna tax team.

The file should tell one coherent story

A completed tax job usually contains several versions of the same underlying facts.

A company profit figure may appear in:

• the trial balance
• financial statements
• tax reconciliation
• taxable income workpaper
• company tax return

A shareholder loan may appear in:

• the balance sheet
• general ledger
• Division 7A workpaper
• prior-year tax file
• loan agreement

A trust distribution may appear in:

• the distribution resolution
• beneficiary accounts
• financial statements
• trust tax return
• beneficiary tax returns

Good review is partly the process of confirming that those documents tell the same story.

When they do not, the reviewer needs language that is more precise than "wrong".

Category 1: information not found

"Not found" means the reviewer expected to find information in the material provided but could not locate it.

Examples:

• no trust distribution resolution in the file
• no franking account workpaper for a company that paid dividends
• no evidence supporting a material capital gain calculation
• no prior-year balance for a carried-forward loss

Not found does not automatically mean the tax treatment is wrong.

The document may exist elsewhere. The fact may be supported by another system. The item may not be required in the particular circumstances.

The correct reviewer response is to ask what the missing information means for the conclusion.

A useful review note is:

"Not found in documents reviewed. Confirm whether the document exists elsewhere and whether it is required to support the treatment adopted."

That is more useful than either passing the item or declaring it incorrect.

Category 2: single source only

Sometimes a figure is found, but only once.

For example, a $240,000 shareholder loan may appear in the balance sheet but there is no loan workpaper, ledger detail or prior-year comparison in the file.

The figure has been found. It has not necessarily been corroborated.

Whether a single source is enough depends on the item.

A bank interest figure from an ATO pre-fill report may not need three separate sources. A complex capital gain or Division 7A loan usually needs a more complete trail.

The reviewer should therefore ask:

• Is this source reliable for the purpose?
• Would we normally expect another document to support it?
• Is the item material?
• Does the figure require a calculation or legal conclusion that is not visible?

"Single source" is a prompt for judgement, not an automatic exception.

Category 3: information agrees across sources

This is the ideal mechanical review outcome.

The same figure or fact appears consistently in the documents where it would reasonably be expected.

Examples:

• taxable income agrees between the tax reconciliation and company return
• beneficiary entitlements agree between the distribution resolution, accounts and trust return
• dividend amounts agree between minutes, statements and the general ledger

Agreement does not prove the legal treatment is correct. It does remove one type of risk: internal inconsistency.

That allows the reviewer to spend more time on the technical conclusion rather than reconstructing the numbers.

Category 4: mismatch

A mismatch means the same item appears in multiple places but the values or facts differ.

Examples:

• taxable income is $410,000 in the workpaper and $401,000 in the return
• a trust resolution distributes 60% to one beneficiary but the accounts show 50%
• the franking workpaper shows a $30,000 debit but the dividend statement implies $28,000
• the opening loan balance does not agree to the prior-year closing balance

A mismatch is stronger than a missing document because there is positive evidence that the file contains conflicting information.

The reviewer should identify:

1. 1the sources that conflict
2. 2the size and nature of the difference
3. 3which source is expected to be authoritative
4. 4whether the difference is a timing issue, rounding issue, amendment or error
5. 5what was changed to resolve it

The final file should not simply overwrite the wrong figure and erase the audit trail. A short note explaining the resolution can be valuable later.

Category 5: conflicting narrative facts

Not every mismatch is numerical.

The client may tell the accountant that a property became a rental in March, while the rental schedule shows expenses from July. A company register may show one shareholder while the current group chart shows another. A loan agreement may name a borrower that differs from the general ledger account.

These matters require the reviewer to determine which fact is correct before applying the tax law.

A technically perfect calculation based on the wrong fact is still wrong.

Category 6: calculation exists but inputs are unsupported

This is a common workpaper problem.

The spreadsheet may calculate perfectly, but the source of the inputs is unclear.

Examples include:

• CGT cost base with no acquisition documents
• depreciation calculation with unsupported opening written-down values
• trust loss workpaper with no evidence for prior-year losses
• franking reconciliation built from manually entered tax payment dates with no ATO account support

A reviewer should separate:

• whether the formula works; and
• whether the inputs are supported.

Both need to be satisfactory for a reliable conclusion.

Category 7: source documents exist but the conclusion is not documented

The opposite problem also occurs.

A file may contain all the relevant documents, but no workpaper explains why the treatment was adopted.

For a simple matter, the documents themselves may be enough. For a complex matter, the Tax Practitioners Board's record-keeping guidance expects records to show the nature, scope and outcome of the service and, for more complex advice, the relevant facts, assumptions and reasoning.

The reviewer should consider whether someone looking at the file later could understand the conclusion without starting again.

Materiality and risk still matter

An incomplete file should not be reviewed as though every missing document has the same importance.

The reviewer should consider:

• dollar materiality
• tax risk
• complexity
• whether the item affects other entities in the group
• whether the issue is recurring
• whether the treatment depends on a legal document
• whether the position is likely to be queried later

A missing $40 receipt and a missing trust deed are both missing documents. They are not the same review problem.

A simple review status model

A useful review system can use statuses such as:

This avoids turning every review result into a false binary. It is the same distinction that sits behind our tax return review checklist for accountants.

Reviewer queries should be specific

Poor query:

"Please check dividends."

Better query:

"The dividend statement records a $21,429 franking credit but the franking workpaper records $18,750. Please confirm the corporate tax rate for imputation purposes and update the workpaper or statement as required."

A specific query tells the preparer what the reviewer saw and reduces the chance of a superficial response.

Close the loop before sign-off

Every material exception should end in one of four outcomes:

1. 1evidence obtained and the item cleared
2. 2calculation or document corrected
3. 3professional judgement recorded and the position accepted
4. 4matter remains unresolved and the file is not ready for sign-off

A review system that identifies exceptions but does not record resolution creates another incomplete file. This connects directly to a firm's quality management system obligations for tax agents.

Where Taxpartna fits

Taxpartna is designed around this distinction between finding information and judging what it means.

It can read across the submitted file, locate relevant information, compare figures and identify where something is missing, supported by only one source or inconsistent across documents. It works as tax workpaper review software that surfaces those results.

Those results are then presented to the practitioner for assessment.

Taxpartna cannot know a fact that was never put on the file. It also cannot decide whether a missing document is legally required in every circumstance. The reviewer remains responsible for deciding whether the evidence is sufficient and whether the tax treatment is ready for sign-off.

Authoritative sources

Record quality expectations for tax practitioners are set out in Tax Practitioners Board guidance. The following primary sources should be checked when this page is technically reviewed or materially updated.

• TPB - Keeping proper client records
• TPB - The use of Artificial Intelligence and the Code of Professional Conduct

Frequently asked questions

Important information. This page provides general information about tax file review and record quality. It does not prescribe a mandatory review method or replace the registered tax practitioner's professional judgement. Taxpartna is a quality assurance assistance platform and does not provide tax advice, prepare returns or give final sign-off.