The binding rulings system was introduced in 1992 to give effect to proposals in a 1991 report, Improvements to Self Assessment Priority Tasks; Information Paper. The decision to make rulings binding on the ATO was driven largely by two circumstances: the introduction of the self-assessment system in 1986 which, it was said, put greater responsibility on taxpayers to ensure positions taken in their income tax returns were correct prior to lodgement, and the long-standing judicial position that, in administering tax laws, the ATO could not be estopped from changing an announced view to the detriment of taxpayers, even retrospectively.
Parliament's response was to legislate a means to hold the ATO to its publicly- and privately-declared positions provided they were expressed in formal ways: in Public Rulings notified in the Gazette, and in Private Rulings issued upon application by affected taxpayers. Taxpayers were specifically allowed to ask for a ruling about a proposed transaction, and the ATO's answer was to be justiciable: taxpayers could test the merits of an unfavourable answer before a Court and secure certainty about the tax consequences prior to doing the deal.
From the outset, the binding rulings system faced difficulties. In the first case on the provisions (CTC Resources) the taxpayer applied for a private ruling on the consequences of a transaction for the 1993 income year but the ATO took so long to respond to the taxpayer's objection to the unfavourable ruling that the 1993 year had ended by the time it replied. The ATO won at trial on the basis that the ATO's delay made its response to the original ruling non-justiciable.
A second problem has become obvious over time. The legislation which gave effect to the rulings system made certain documents issued by the ATO binding on the organisation, but it did not require the organisation to issue documents; it merely said that documents, if issued, would be binding once they met certain criteria. The evidence of the last two decades is that the ATO has become reluctant to issue, of its own volition, the types of public documents offering general interpretation of tax legislation which can become binding on it. For example,
- in 1997, the ATO issued 25 public rulings in the TR series;
- by 2007, that number declined to 13;
- in 2017, there were only 5 (3 TRs and 2 LCRs, equivalent to TR but meant to explain new legislation).
Over that period, the ATO has chosen to create classes of non-binding documents such as Practical Compliance Guidelines, ATO Interpretive Decisions and Law Administration Practice Statements. It also shifted to providing guidance on specific transactions through Class and Product Rulings, which are effectively a hybrid of a public and private ruling.
A third problem has also become apparent in the strict way that rulings are viewed. For example, the taxpayers in the Resource Capital Fund litigation could not hold the ATO to its ruling that article 7 of the US treaty denied Australia the ability to tax the foreign investors if the ATO decided to proceed against them under article 13 instead. The taxpayer in Bellinz discovered that rulings denying the user of a fixture the entitlement to depreciation deductions did not carry the implication that the financier became entitled to the depreciation instead. The taxpayer in Pratt Holdings discovered that a ruling issued to the transferor of losses could not be relied upon by the transferee of the losses. And the taxpayer in ABB found it was not entitled, in the context of a withholding tax dispute, to rely on the ATO's ruling on the time of derivation for income tax. While all of these positions are in varying degrees logical implications of ATO rulings, the ATO did not like these implications and the Court permitted this, insisting on a very precise compass for the Rulings.
And a final lesson is seen in the recent decision in Hacon. While the ATO has complete control over whether to initiate a document and the form it will take, it always had rather less control over requests which come to it from taxpayers. For example, taxpayers initiate the requests for the ATO to issue Product Rulings, Private Rulings and Class Rulings, but the ATO's handling of requests for Private Rulings has been contentious. The impression emerged that the ATO was reluctant to issue private rulings which might become binding on it. Taxpayers who asked for a formal ruling were sometimes offered other things instead: an informal `letter of comfort' or an assurance that `no compliance resources would be allocated' to pursuing a point. And some topics appeared to be `no go areas.' For example, at various times the ATO has been reluctant to give private rulings on Part IVA, the meaning of `fixed trust' or `bare trust,' whether the `same business test' would be met, and the operation of s. 974-80.
The Hacon case is an example of a `no go area,' compounded by the ATO's extra reluctance to rule on future transactions. The transaction involved the proposed issue of dividend access shares and the application of Part IVA to that transaction. Understandably, the taxpayer did not want to proceed with the transaction without the certainty of a ruling. The ATO declined to issue a Private Ruling, even after the taxpayer volunteered some further information, and then offered to supply any `further details of the facts and circumstances [the ATO might require] to issue a private ruling ...'
The question behind Hacon whether the ATO is obliged to rule on every ruling request it receives is not a simple one, and the legislation tries to navigate a course between the value of giving taxpayers certainty through a ruling and the need to protect the ATO from having to perform unnecessary work, especially in respect of proposed transactions. The balancing of these competing interests can be seen in a number of design features which were present from the first version of the legislation and remain, albeit expressed rather differently, in the current version.
Conceptually, the private rulings regime might have handled this tension in a number of ways:
1. the ATO would always be obliged to issue a ruling (unless the application was frivolous or vexatious or entirely speculative) and to rule based only on the information in the application. The taxpayer would carry the risk that the ruling would turn out to be worthless because the information in the application was deficient or incorrect); 2. the ATO would not be obliged to issue a private ruling if it judged that all the relevant information has not been supplied in the application, but in that case:
- the ATO was then obliged to ask for the missing information and once the information was provided, the ATO was obliged to issue the ruling; or
- the ATO was not required to seek more information; instead it was obliged or entitled to make the assumptions it thought most appropriate and then issue the ruling; or
- the ATO was not required to seek more information or make assumptions; it had a discretion not to issue the ruling; or
3. the ATO would be given an absolute discretion whether to issue a private ruling, and in all cases.
The problem in Hacon is that the legislative design is ambiguous. Section 359-5 says, `the Commissioner may ... make a written ruling ...' while s. 359-35 says, `the Commissioner must ... make the ruling...' They both can't be right. And while there are some exceptions to s. 359-35 they arise in specific circumstances: `making the ruling would prejudice or unduly restrict the administration of a taxation law' or `the matter ... has [already] been considered by the Commissioner' or if the taxpayer wants a ruling on the exercise of a discretionary power.
And the design is made even murkier by the rules dealing with the situation where the ATO feels the information in the application is inadequate for issuing a ruling. Section 357-105 `the Commissioner must request the applicant to give [the missing] information ...' while s. 357-110 says the Commissioner may, `make such of the assumptions as the Commissioner considers to be most appropriate' but the ATO also has the discretion to `decline to make the ruling.' The obvious question is whether the ATO can decline to make the ruling without first asking for more information, which is what happened in Hacon.
At first instance, the Federal Court said, `the Commissioner [was] subject to an imperative obligation to request information which he considered necessary ...' but on appeal the approach was quite different. The Full Court accepted the ATO's position that, no matter how much further `information' the applicant might provide, at the end of the day the Commissioner would still `have to make assumptions about future events' and so he was permitted not to rule. And the Court permitted the ATO to come to that conclusion a little prematurely: it could decide that assumptions would always be needed, without first asking the applicant for more information to see what that further information might reveal.
The ATO has recently issued a Decision Impact Statement about Hacon which says, `the decision confirms [the ATO] is entitled to decline to make a private ruling where the correctness of the ruling would depend on assumptions about future events or other matters, and that the Commissioner is not obligated to first request that information from the taxpayer in those circumstances.'
This is an accurate summary of the result of the case but the judgment reveals a very significant limitation on the ability to get private rulings about proposed transactions. In some circumstances it will be true that the proper analysis of today's transaction will be coloured by what tomorrow's state of affairs turns out to be, and tomorrow's state of affairs might be very fluid indeed. But many Private Ruling applications about a proposed transaction necessarily involve an uncertain tomorrow. This judgment invites the ATO to ponder tomorrow's world, to speculate about its uncertainty, and to conclude that its unpredictability offers an adequate reason not to rule.
The more sensible answer to the problem in Hacon is to insist the ATO must always rule unless an application is frivolous or vexatious (as s. 359-35 already provides). The system might insist that the ATO seek further information (as s. 357-105 already provides). If the ATO believes that assumptions will still have to be made, the legislation should not allow that to be an escape route; rather, the legislation would require the ATO to ask the taxpayer what assumptions the taxpayer wants the ATO to make about tomorrow's world (as s. 357-110(2) envisages). The ATO must then rule. If the ruling is favourable but it was based on misinformation or assumptions which are not borne out in tomorrow's state of affairs, the ruling is worthless. And if the taxpayer chooses to challenge an unfavourable ruling, a favourable judgment based on incorrect information or unmet assumptions would be equally worthless.
This sets exactly the right incentives for a taxpayer who wants a reliable Private Ruling: to deliver full and accurate information, to describe assumptions with as much foresight as it can muster, and to ensure as best it can that tomorrow's world comes to match those assumptions. It places the risk on the taxpayer, and ensures the ATO delivers the benefits the binding rulings system was meant to offer. The ATO needs to be protected against having to rule on applications that are frivolous, vexatious or entirely speculative. But that was not the situation in Hacon; the transaction was definitely in serious contemplation. The Court's judgment allows uncertainty about the future to be the ATO's escape route; uncertainty about tomorrow should be a risk which the taxpayer bears and has to manage as best it can.
In fact, the Court in Hacon might have read the current legislation in exactly that way the constituent elements are there already. The only dilemma for the Court was to assemble the components in the right sequence. Instead, the interpretation adopted in Hacon arguably undermines the original policy behind the rulings system: to give taxpayers certainty and in advance, even for future transactions, and even when the final situation is still in a state of flux.