Sub-trust interest, the assumption that was not tested, and the onus problem
1. Cameron v Commissioner of Taxation1 decides two things: interest on a sub-trust loan was not deductible under s 8-1 because the supposed borrowing was not productive of the Family Trust’s assessable income; and that the Applicants did not discharge their onus that no family trust election had been made.
II. The deductibility holding
A. The arrangement
2. Adjuvant Pty Ltd was a corporate beneficiary of the Cameron Family Trust. Acropolis Pty Ltd was trustee of the Family Trust. Adjuvant’s entitlements to Family Trust income were set aside on what the parties called the Adjuvant sub-trust, of which Acropolis was also trustee. The sub-trust amounts were then said to have been lent back to the Family Trust on Option 1 terms under PS LA 2010/4.2
3. No money moved. Hill J describes the mechanic at [11]:
The amount to which the Family Trust was entitled under the sub-trust borrowing for a year was set off against Adjuvant’s unpaid present entitlement from the Family Trust for that year. That is, no money was actually advanced by the Adjuvant sub-trust to the Family Trust.
4. The mechanic is set-off, not advance. The Family Trust’s liability to pay Adjuvant’s UPE was extinguished against an equal amount said to be owing under a loan from the sub-trust. The Family Trust paid annual interest on that supposed loan and claimed the interest as a s 8-1 deduction.3
B. The use-of-funds nexus
5. Hill J’s analytical move is at [59]. He frames the question as whether the supposed borrowing was put to a use that was productive of the Family Trust’s assessable income.
Here, the Sub-trust Loans were not used to produce income for the Family Trust directly. As noted, the amounts lent by Adjuvant under the Sub-Trust Loans were set off against UPEs held by Adjuvant in respect of the Family Trust. It was common ground that the direct effect of these arrangements from the Family Trust’s perspective was to convert one type of liability into another: the Trust’s liability to pay Adjuvant’s UPE (an immediate liability, not bearing interest) was converted into a debt liability (a non-current liability to repay the principal in seven years, and to pay interest in the meantime).
6. The reasoning is conventional. Interest is deductible because of the use to which the borrowed money is put. If the borrowing buys an income-producing asset, the interest follows the asset. If the borrowing replaces something, the interest follows what was replaced. Here, on Hill J’s analysis, the borrowing replaced nothing income-producing. It replaced a non-interest-bearing UPE with an interest-bearing loan. The Family Trust acquired no new asset and lost no productive asset. The set-off was an accounting event.4
C. The Total Holdings argument
7. The Applicants ran an indirect-productivity argument. By not calling on the Camco 3 Trust for payment of distributions, the Family Trust left funds available to the Camco business; the Camco business was therefore more profitable; future distributions to the Camco 3 Trust would be larger; and future distributions through to the Family Trust would be larger. The argument was framed by reference to Total Holdings.5
8. Hill J rejects it at [65]:
legal form is a relevant part of the circumstances (even if it is not determinative), and it is not sufficient for a taxpayer to show that the outgoing is similar or economically equivalent to an expenditure that might have been incurred … Choosing not to call on a present entitlement under a trust is very different from purchasing shares and entering into loans.
9. The point is sound. Total Holdings was a case where the taxpayer borrowed externally and on-lent to a subsidiary. The borrowing produced something. Here the Family Trust did nothing other than not call on a UPE that it could not have collected against in any event because the Camco 3 Trust did not have the cash. The Family Trust took no positive step that could have a productive object.6
10. Hill J adds the obvious extension at [67]: if the argument were accepted, the Family Trust’s borrowings would generate deductible interest even if used for purposes entirely unconnected with the Camco business, on the footing that not calling on the UPE preserves funds in Camco 3 Trust. That cannot be right. The connection between the supposed interest outgoing and the gaining of assessable income is, on his Honour’s assessment, too remote.7
D. The Roberts argument
11. The fallback was Roberts. The Applicants said that the Sub-trust Loans preserved capital in the Camco 3 Trust, and that Roberts supports the deductibility of interest on a borrowing used to preserve capital in a productive enterprise.8
12. Hill J reads Roberts narrowly at [75]:
Roberts does not assist the Applicants, for two separate reasons. First, in no sense has the Family Trust invested in the Camco 3 Trust, let alone in a manner comparable to partnership capital contributions. Second, on the evidence, the Sub-trust Loans are not entered into to preserve the capital contributions made by the Principals: rather, Mr Cameron refers to the Sub-trust Loans ensuring that the Camco business “retained as much as possible of the funds representing its profits”.
13. The reading is correct. Roberts is about replacing partnership capital contributions, not retaining undrawn profits. The Family Trust never made a capital contribution to the Camco 3 Trust. The thing being preserved was, on Mr Cameron’s own evidence, retained profits, not contributed capital.9
E. The alter-ego problem
14. Beneath both arguments lay a single fault. The Applicants treated the Camco entities and the Family Trust as a single economic enterprise. Hill J rejects that at [66]:
it is not open to ignore the different structures in place, and to treat the Camco Trusts (such as the Camco 3 Trust) and the various family trusts (such as the Family Trust) as an aggregate entity with aggregate activities.
15. Trusts in the same family group are not one entity. They derive income separately. They are taxed separately. They have separate trust deeds, separate trustees and separate beneficiaries. Submissions that ignore that fact fail.10
III. The assumption that was not tested
A. The sub-trust was assumed to exist
16. Throughout the deductibility analysis Hill J takes as a given that the Adjuvant sub-trust existed, that the Sub-trust Loans were loans, and that the Family Trust had incurred interest. None of those characterisations is examined. The analysis runs entirely at the level of whether the supposed borrowing was put to a productive use.
17. The pleadings explain it. The Commissioner’s objection decisions in the deductibility appeals are dated 17 October 2023; the AAT decision in Bendel was handed down on 28 September 2023. Whatever the Commissioner now thinks of the trust-law characterisation of PS LA 2010/4 sub-trusts, the case had been framed on a different theory.11
18. Hill J at [77] expressly leaves the Commissioner’s other arguments to one side:
It is not necessary to rule on the Commissioner’s other arguments as to why the Interest Payments are not deductible; for example, the Commissioner contended a separate basis for refusing deductibility was that the Family Trust, as the beneficiary of a discretionary trust, never had an entitlement to receive any money from the Camco 3 Trust, and its receipt of income from the Camco 3 Trust was passive and did not involve any “activities or operations”.
19. Cameron decides what was pleaded. It does not decide what was not. The deeper question, whether the sub-trust existed at all, was not put.12
B. Income is not property: the real Bendel point
20. In my view a PS LA 2010/4 Option 1 sub-trust of the kind the parties used in Cameron is not a valid sub-trust. A trust requires trust property. Trust income, by itself, is not property susceptible to a separate trust. A resolution that purports to settle a beneficiary’s entitlement to a share of trust income on a further trust does not constitute that further trust because there is no separately identifiable property to be its subject matter.
21. The Tribunal at first instance in Bendel13 put the proposition directly:
Income is not property. If any property is created by the resolution … it is the property of the beneficiary, namely the right to be paid.
22. That holding does the work that needs to be done. A resolution that the trustee will pay a beneficiary a share of the income of the trust creates, on the side of the beneficiary, a right to be paid. It does not carve out trust property and put it under a fresh trust. The trustee continues to hold the assets of the head trust on the head trust’s terms. The beneficiary acquires no proprietary interest in any segregated fund.14
23. The doctrinal foundation is the divergence in Ward. North P and McCarthy J held that an unconditional allocation can vest an equitable interest in money already in hand. Turner J disagreed: no new trust arises absent identification of the specific property impressed with it. The distinction is between equitable vesting (the beneficiary acquires a right enforceable against the head trustee) and resettlement (the head trustee constitutes a new trust over identified property). The High Court in Fischer v Nemeske confirmed that application can occur by irrevocable allocation alone, without resettlement.15
24. Applied to the Adjuvant arrangement, the resolution to set aside Adjuvant’s entitlement on sub-trust does what the McCarthy J / North P line in Ward permits: it confirms an equitable right in Adjuvant against the head trustee. It does not do what Turner J in Ward would require for a fresh trust to arise: identify property and impress a new trust on it. Acropolis continued to hold the head trust assets on the head trust’s terms. There was no segregated fund. There was no new corpus. There was no sub-trust.
25. The High Court’s questioning in Bendel reinforces the position. The bench tested the proposition that a trustee’s resolution and accounting entry can produce a debtor-creditor relationship. Gordon J’s position, in particular, was that until the beneficiary calls under Saunders v Vautier, the equitable relationship endures and no unconditional present obligation to pay arises. Edelman J added that an outstanding trustee’s indemnity defeats any Saunders v Vautier call. The implication is that present entitlement in the Harmer sense, requiring the immediate right to demand and receive payment, may not arise on the mere passing of a resolution.16
26. That reasoning, if accepted, is fatal to PS LA 2010/4 Option 1 in two ways. First, there is no trust property to be the subject of the supposed sub-trust; income is a flow, not a thing. Second, the corporate beneficiary is not presently entitled to anything segregated and demandable; the seven-year Option 1 architecture is precisely a structure that prevents present entitlement, because the beneficiary cannot call for immediate payment.
27. If the sub-trust did not exist, three things follow. There was no sub-trustee holding identifiable funds capable of being lent. There was no loan. There was no outgoing properly characterised as interest within s 8-1. The s 8-1 question Hill J decides never arises.
C. The narrowness of the case
28. Cameron is therefore authority on a particular question: whether interest on a PS LA 2010/4 set-off arrangement, taken at face value, satisfies the use-of-funds nexus under s 8-1. The answer is no. The case is not authority on whether the arrangement is what it says it is. That question is alive, and is before the High Court.
IV. The onus problem
A. What the taxpayers had to prove
29. The second limb of Cameron concerned family trust distribution tax on distributions outside the family group. The Applicants’ case was that no family trust election had ever been made. The Family Trust’s 1999 income tax return, and the 1999 returns of five related entities, said that one had. The Applicants had the onus.17
30. The Applicants offered two propositions. First, that the firm’s tax agent, Mr Cook, had never completed the 1998 ATO form. Second, that the election data in the 1999 returns was draft or template information accidentally left in. Mr Cameron’s evidence was that he had never heard of a family trust election.
31. Hill J held that the Applicants had not discharged the onus. The case is a clean worked example of how an onus dispute is lost when the documentary record cuts against the witness recollection.18
B. The methodology
32. Hill J’s methodology is the Fox v Percy / Gestmin methodology applied to a 25-year-old dispute. He records it at [109]:
These considerations have encouraged judges “to limit their reliance on the appearances of witnesses and to reason to their conclusions, as far as possible, on the basis of contemporary materials, objectively established facts and the apparent logic of events”.
33. The proposition is straightforward. Where events happened many years ago, contemporary documents are a safer guide than later recollection. Confidence in a recollection is not a reliable indicator of its accuracy. A witness can be honest and wrong.19
34. Mr Cook’s evidence was 25 years out from the events in question. Mr Cameron’s evidence was that he had no recollection of a concept he would have had to have understood if his witness account were correct. Hill J declines to prefer either witness over the documentary record. He puts the conclusion at [128]:
Considering the evidence as a whole, I would not accept the recollections of Mr Cook and Mr Cameron as to whether or not the 1998 ATO form was ever signed. The events occurred more than 25 years ago. This is exactly the situation to which the comments in Fox v Percy and Gestmin are directed.
35. That is the right approach to the evidence. A 1999 income tax return saying an election was made is the contemporary record; a 2024 affidavit saying it was not is not.20
C. Equipoise is failure
36. On the documentary record itself, the position was finely balanced. The 1999 returns suggested an election had been made; the 2000-2022 returns suggested it had not. Hill J resolves the tension by reference to the onus at [131]:
Here, the evidence before the Court does not allow me to be satisfied that it is more probable than not that a family trust election was not made in respect of the Family Trust. Putting the matter most favourably to the Applicants, the evidence gives rise to competing inferences of equal probability.
37. The taxpayer must persuade. The civil standard is not a comparison of probabilities; it is a requirement of actual satisfaction. Where the evidence is in balance, the onus has not been discharged. The taxpayer loses on equipoise. That is the orthodox tax onus position; Cameron is its application to an old-event dispute.21
D. The practical lesson
38. A taxpayer running an onus case on stale facts should not rely on witness recollection to displace the contemporary documents. Either the documents support the taxpayer, in which case the recollection is unnecessary; or they do not, in which case the recollection will not save the case. The earlier the taxpayer collates the documentary record, the better positioned the taxpayer will be when the dispute crystallises.
39. The proposition that detailed election data was accidentally left in six related-entity returns, by the same agent, lodged for two different client groups at different times, is implausible on its face. Once Mr Cook’s explanation of how that came about was internally inconsistent (template error in one affidavit; deliberate-but-forgotten draft data in cross-examination), the case was lost. The reconstruction problem was not the taxpayer’s fault. It was the consequence of trying to prove a negative across 25 years.
V. Conclusion
40. Cameron decides what was put to Hill J. On s 8-1, taken at face value; the arrangement fails because the set-off produces nothing income-producing. On family trust distribution tax, the taxpayer’s recollection of events 25 years past does not displace contemporary documents that point the other way.
41. The case is not authority on whether the Adjuvant sub-trust existed. The Commissioner did not run the point. The first-instance reasoning in Bendel that income is not property, and the High Court’s questioning around Saunders v Vautier and the trustee’s indemnity, make it doubtful that a PS LA 2010/4 Option 1 arrangement produces either a sub-trust over identifiable trust property or a present entitlement in the corporate beneficiary. The High Court will decide what is left of the architecture. Practitioners advising on existing PS LA 2010/4 arrangements should not read Cameron as comfort that those arrangements work subject only to a s 8-1 use-of-funds point. They may not work at all.