Melbourne University Law Review
[For almost three decades, the Australian Treasury has issued an annual ‘tax expenditure statement’ detailing concessions in Australia’s tax laws. It was originally argued that tax expenditure budgets (the international term for these statements) would lead to simpler laws with fewer and better targeted concessions. This clearly has not happened in Australia, as tax laws have become more complex and tax concessions less efficient since the tax expenditure budget concept was accepted in Australia. The problem is not with the concept itself but rather with its execution. In particular, Australia has copied a United States model — a model that may be very inappropriate in the context of Australian tax jurisprudence. In Australia, many tax concessions that give rise to uncertainty and complexity have not been introduced explicitly by the legislature but rather result from judicial doctrines that have been implicitly endorsed in the design of the tax law. Tax expenditure analysis will not yield better outcomes until it is extended to these implicit tax expenditures.]
More than 40 years after the Assistant Secretary of the United States Treasury, Stanley Surrey, first coined the phrase and articulated the concept and almost three decades since the first Australian tax expenditure statement was prepared, tax expenditure budgets have become a key tool for analysing government revenue collection and spending policy.
Tax expenditure budgets record the fiscal cost of deliberate deviations from a neutral tax system, particularly measures providing concessional treatment for particular taxpayers or activities, and, in some cases, analyse as well their economic effect and their impact on the fairness of the tax system. The budgets measure the cost of concessions by calculating the tax that would have been collected under a neutral ‘benchmark’ tax law without any concessions and then treating the value of taxes actually forgone as a result of legislated concessions as substitutes for equal value cash grants. It is the equivalence of taxes forgone as a consequence of tax concessions and the direct expenditure that could have been made to provide similar fiscal benefits that has led to the term ‘tax expenditure’ for these deviations from a neutral tax base. In Australia, the term (and some might claim the concept) has been extended to the opposite situation, where additional tax is imposed on taxpayers relative to the neutral benchmark because they are denied deductions for non-personal expenses incurred to derive assessable income or are required to defer recognition of those expenses. Measures leading to overtaxation of this sort are sometimes labelled ‘tax disincentives’.
In a sense, the tax expenditure concept is merely a restatement of a fundamental premise of tax policy: taxes should raise revenues needed for public goods and services and to achieve the redistribution goals of social equity in a way that leaves the smallest economic footprint and causes the least interference possible in the operation of the market. This is not to say that governments should not use tax legislation to achieve social or economic goals. Often policymakers deliberately intervene in the market to promote transactions that yield positive externalities, to discourage those that generate negative externalities, or to overcome market failures. While many vehicles apart from tax rules can be used to address these issues, embedding needed subsidies or disincentives in a tax law rather than another law is as legitimate an option as all others for policymakers. Tax expenditure analysis simply states that spending programs and disincentives inserted into tax laws should be subject to the same level of rigorous scrutiny as parallel programs outside the tax legislation to determine whether the intervention mechanisms achieve their goals in an efficient and equitable manner, with the least collateral distortion possible.
Badly designed tax expenditures have been criticised in their own right as inefficient subsidy programs and separately for the complexity they bring to the tax law and the compliance costs they impose on taxpayers and tax administrators. The process of subjecting tax expenditures to scrutiny through exposure in tax expenditure budgets has, to some extent, led to reductions in the inefficiencies or inequities of some concessions. Some upside down concessional deductions or exemptions have been replaced by disappearing and refundable credits and others have been replaced by more efficient and better targeted direct spending programs. Exposure and critical review of tax expenditures has, however, had no apparent impact on the overall complexity of the law or consequent compliance costs. To the contrary, in the decades since tax expenditure analysis was incorporated into Australian budget processes, tax laws have become increasingly complex and unmanageable, and consequential compliance costs continue to escalate. In short, as a tool of reform, tax expenditure identification, exposure and analysis appears to have been pretty much a failure.
The inability of the tax expenditure concept to translate into better tax laws is not a consequence of inherent weaknesses with the model but rather results from the way in which it has been used or, more significantly, not used. Tax expenditure analysis has failed to lead to simpler or more efficient tax legislation because the concept has been used at the wrong time and in the wrong way, and applied to the wrong target. It has usually been applied retrospectively, after law is made; it has been used to evaluate the operation of laws rather than guide their design; and, most significantly, it has been limited to explicit legislative measures that affect the calculation of taxable income, ignoring the implicit tax expenditures that shape much of the tax system and generate many of its most significant distortions. Until these failings are addressed, tax expenditure analysis will remain an impotent tool for tax reform.
The starting point for Surrey’s tax expenditure analysis was the tax legislation. In the case of income concessions, identification of tax expenditures was merely a matter of scanning the legislation for full or partial exemptions or exclusions. In the case of deduction concessions, it was a matter of running through the allowable deductions and identifying instances when taxpayers were allowed deductions for personal consumption expenses rather than inputs to derive gross income. A third set of timing tax expenditures could be identified by noting statutory measures that allowed taxpayers to defer recognition of income until later years or accelerate recognition of expenses before they were fully incurred in an economic sense. A fourth set of tax expenditures was included in the law as tax credits (now known in Australia as ‘offsets’) that reduced and in some cases refunded taxes otherwise due to subsidise a range of transactions.
The Australian Treasury’s wholesale importation of the United States (‘US’) approach of looking only at the tax legislation for tax expenditures is not surprising. At first glance, the legislation in the US and in Australia look similar. The income tax laws in the US and in Australia impose tax on a measurement of net gains known as taxable income and in both cases the calculation of this net amount starts with the inclusion of ‘income’ followed by deductions for expenses incurred to derive the income. In both cases expenses are categorised as capital or non-capital outlays and in both cases there are rules that allow the immediate deduction of non-capital expenses and the deduction over time of capital expenses through a depreciation or capital allowances regime. On paper, at least, the laws look pretty much the same, and starting with a proven operational model added legitimacy to the arguments of those advocating adoption in Australia of tax expenditure budgets in the face of reluctant or even hostile politicians.
A closer examination reveals, however, that the superficial language and structural similarities between US tax law and its Australian counterpart mask significant differences in the conceptual foundations of the two laws. Both laws may start with income, for example, but there is no shared understanding of what constitutes income. In the US judicial concept of income, a buck is truly a buck. The calculation of taxable income commences with ‘gross income’ and US courts have interpreted the term to include virtually all receipts, whatever their form or character for other purposes. Payments in respect of labour, business and property are income but so, too, are complete windfalls — money found purely by chance or unexpected gifts. Receipts and gains are removed from the tax base only by explicit legislative intervention, and concessional treatment of particular types of receipts can be identified simply by looking in the legislation for full or partial exclusion rules or deferred inclusion measures.
The Australian legislation starts with the same term, ‘income’, but the judiciary in this jurisdiction, using the ‘transplanted categories’ and ‘transplanted outcomes’ doctrines, has read down the income concept so it captures only a small slice of receipts that would be considered income by US courts. The transplanted categories doctrine refers to the practice of importing into tax law concepts from other areas of law, such as the use by Australian judges of trust law tests to define income for tax purposes. Trust law dissects gains derived by a trustee into two categories: income gains, to which life or income trust beneficiaries are entitled, and capital gains, to which remainder or capital beneficiaries are entitled. Applying trust law tests to income tax, the judiciary narrowed the concept of income for tax purposes so it included only amounts that would have an income character for trust law purposes — receipts that are periodic in nature, anticipated by the recipient, in the form of cash or something convertible to cash, realised and severed from the source, and paid in respect of the provision of services, the use of property, or a business transaction.
The transplanted outcomes phenomenon reflects what is sometimes referred to as the ‘colonial cringe’ mentality of Australian judges in the first decades (or longer) of independence — an assumption that precedents of United Kingdom (‘UK’) judges interpreting UK law should govern the interpretation in Australia of Australian law, with little regard to differences in the legislative regimes. The UK schedular income tax was fundamentally different in structure from the income tax laws enacted in the Australian colonies prior to Federation and the income tax Act adopted by the Commonwealth during the First World War. Australian legislators deliberately eschewed the UK system of classifying receipts and outgoings into ‘schedules’ and then ‘cases’ within schedules in favour of global income taxes, but when it came to interpret the scope of the global income tax, Australian judges looked to English precedents — even first instance decisions — as guides to the meaning of terms in Australian law. As a result, a receipt that was found by a UK court to be outside the scope of a narrow case within a narrow schedule in the UK statute would be assumed to be a capital gain that fell outside the ‘income’ concept in Australia. In the post-Second World War period, judges in Australia gradually shifted from blind adherence to English precedents, but by then the consequences of colonial cringe had been embedded in local cases that would become the precedents for the next generation of cases.
Australian judicial concepts also diverged from their US counterparts significantly on the deduction side. Like the US law, the Australian tax legislation denied an immediate deduction for capital expenses and contained separate rules for the later recognition of these expenses. Australian courts took a very different approach to determining what constituted a capital outlay, however. The US judicial concept by and large characterises expenses as capital outgoings when they yield benefits that last well past the end of the tax year. In Australia, a transplanted distinction between revenue and capital expenses gave only passing consideration to the longevity of an acquired benefit relative to the tax year in which it was acquired. Instead, it looked at the factors originally used to allocate the expenses of a trust to the account of a remainderperson (the capital beneficiary) as opposed to those allocated to the account of a life beneficiary (the income beneficiary). As a result, under Australian tax law the character of an outgoing as a revenue or capital expense depends on factors such as the form of the payment or payments (lump sum or periodic), how often the taxpayer acquires similar benefits or assets, whether the expenditure yields an ‘enduring benefit’ (the judicial test closest in effect to accounting principles) and, most importantly under current judicial tests, whether the expense is said to relate to the ‘process’ of the taxpayer’s enterprise or to its ‘structure’. These Australian judicial doctrines sometimes have the effect of classifying what would be considered revenue expenses in tax policy terms as capital outgoings and vice versa.
Although the US and Australia started with not dissimilar legislative regimes, the remarkably different directions taken by the judiciaries in the two jurisdictions has significant implications for the efficacy of a tax expenditure budget based only on deliberate statutory concessions. The restricted tax expenditure analysis may work well in the US, where judicial doctrines reinforce the underlying tax law and deviations from the benchmark tax base are for the most part limited to explicit concessions inserted in the law. It is woefully inadequate in a country in which judicial interpretations initially emasculated the law and subsequent incorporation into administrative practices and acceptance by the legislature of judicial characterisation then camouflaged a wide range of deviations from the benchmark in practice.
The phenomenon of legislative endorsement of a judicial characterisation can be illustrated with three examples: the treatment of expenses incurred to protect a taxpayer’s legal title to an asset, the characterisation of payments received by individuals in return for non-competition covenants, and the treatment of payments to acquire wasting contractual rights such as a tied-house agreement.
The subject of the first example, an expense to resist a challenge to legal title, is assumed by accounting standards and tax systems abroad to have no ongoing value to a taxpayer (the benefit expires when the challenge is met, whatever the outcome) and is thus normally a deductible expense. Australian courts originally followed this approach. However, following his appointment to the most senior position on the High Court, Dixon CJ was able to elevate his preferred test for distinguishing revenue and capital expenses (the ‘process’ or ‘structure’ test noted earlier) from its initial presentation in a non-majority judgment to become the law of the land and as a consequence, these expenses, formerly treated as revenue outgoings, acquired a capital character in Australian jurisprudence.
A quarter century after expenses incurred to defend legal title were characterised as capital outlays, the legislature finally conceded taxpayers should be allowed to recognise these costs for tax purposes. However, rather than restore the deductibility of these expenses, the drafters of the amending legislation accepted the judicial characterisation of the expenses as capital outlays and provided for recognition of the expense through the capital gains regime. Taxpayers are allowed to add defence of title expenses to the cost base of assets to which they relate, so the outgoings can only be recognised when there is a disposal of the asset. From a benchmark tax base perspective, this treatment amounts to a significant tax penalty. Because the tax expenditure statement does not recognise instances where judicial characterisation deviates from the benchmark tax system, the rules are not included in the tax expenditure accounts.
The subject of the second example, the characterisation of payments received by individuals for negative covenants or non-competition agreements, is a useful example of the transplanted outcome or colonial cringe phenomenon. UK courts, working with the narrow UK tax legislation catching particular types of income that fell within specific cases under defined schedules, had concluded that paying an employee or contractor an amount for not working for a competitor as well as an amount for working for the employer or principal was not taxable, as it fell out of the case that assessed a profit or gain ‘arising or accruing from the taxpayer’s profession or vocation’. While there were no Australian cases directly on point, the Australian Taxation Office (‘ATO’) and Australian Treasury assumed that the payments would be capital payments in Australia following the UK decision based on fundamentally different legislation.
Subsequently, legislative drafters, adopting the assumption of the ATO and Treasury, went out of their way to ensure the receipts were not taxed under Australia’s income tax regimes for fringe benefits or the rules for payments related to retirement or termination of office. They were only brought into the tax net through the capital gains tax regime adopted in 1986, with effect from 1985. In this case, while the legislators retained the judicial characterisation of the receipt in their statutory response, the tax inclusion mechanism removes any preference for this type of remuneration. The payments are treated as realised capital gains, but are excluded from the generous 50 per cent exemption concession provided for most other types of capital gain. They are, as a result, effectively taxed in the same manner as they would be if they were considered ordinary income subject to the general assessment formula. Once again, the tax expenditure statement does not mention the judicial characterisation of these payments, although in this case, since the statutory treatment yields the same outcome as would a benchmark income tax characterisation, there is no deviation from the benchmark and consequently no tax expenditure.
A tied-house agreement, the basis for the third example of tax expenditure analysis failing to capture measures based on judicial characterisations, is a contractual obligation by a retailer to stock exclusively the wares of a manufacturer or to give prominence to that manufacturer’s products. Under a neutral benchmark income tax (and accounting principles), expenses incurred to secure long-term contractual rights such as a tied-house agreement are amortised over the life of the contracts. But at the time when Australian courts were first asked to characterise these expenses, the amortisation rules in the Australian income tax law only applied to expenses for tangible plant and articles. No depreciation was allowed for intellectual property assets such as copyrights or patents or for intangible assets such as multi-year contracts.
The only options open to the courts were to characterise the expenses incurred to acquire multi-year contracts as revenue expenses, allowing an immediate deduction, or as capital outgoings, denying the taxpayers any recognition for their expense. Both options were wrong from a tax policy perspective, with a revenue characterisation amounting to a tax concession and a capital characterisation being a tax penalty. The courts recognised that both options were flawed, but having no other options made the best of a bad situation by characterising expenses for some relatively shorter-term contracts as revenue expenses while treating the costs of acquiring some relatively longer-term contracts as non-deductible capital outgoings.
In the mid 1980s, the legislature finally intervened. It accepted the judicial characterisation of expenditures to secure short-term contracts as revenue outgoings and the characterisation of expenditures to obtain long-term contractual benefits as capital expenses, but adopted rules to modify the tax treatment of both types of expenses. Acquisition costs enjoying a revenue character were made subject to an amortisation regime that requires taxpayers to spread recognition of the cost over the life of the contracts. The amortisation rules were limited to revenue expenses, which automatically excluded the cost of acquiring long-term contractual arrangements that were (correctly) regarded as capital outgoings under judicial tests. These outgoings were incorporated in the capital gains regime which allows taxpayers to recognise the expenses — but only when the contracts expire and then only as capital losses, meaning they cannot be utilised unless the taxpayer has capital gains against which the losses can be offset. As with the previous two examples, the deviation from the benchmark based on judicial characterisation is ignored under the current tax expenditure accounting system.
Implicit tax expenditures and flawed responses to these unanticipated concessions and disincentives have generated much of the complexity, inefficiency and inequity in Australian tax laws. The case for extending tax expenditure analysis to implicit tax expenditures, and particularly to legislative responses that effectively endorse judicial characterisations at odds with benchmark principles, seems compelling. There are, however, a number of technical issues sometimes raised as pragmatic barriers to the application of the tax expenditure concept to implicit expenditures and disincentives. However, these four technical issues — difficulty with costing these deviations, the risk of perceived criticism, the unclear boundaries of the benchmark, and the difficulty in identifying possible objectives for the implicit expenditures — may also apply to explicit tax expenditures.
The principal technical objection to expansion of the tax expenditure budget to include implicit tax expenditures is the difficulty or even impossibility of costing most implicit spending programs or tax disincentives. Many tax expenditures are difficult to cost because the tax administration process does not generate data specific to particular measures. For example, if judicial doctrines allow an immediate deduction for an expense that would be capitalised and amortised over its life in the benchmark income tax, the taxpayer simply includes the expenditure in gross deductions, without flagging that it results from an implicit tax expenditure. A similar problem can arise with a statutory provision, however. If a concessional statutory provision allows the immediate deduction of an expense that could be considered a capital outgoing in a benchmark income tax, taxpayers will include the outgoing in their calculation of deductions and there will be no means of collecting data on the impact of the measure.
Completely accurate costing data is only possible where tax expenditures or disincentives are identified explicitly and corresponding deductions, inclusions, exclusions and so forth are entered separately on tax returns. Accurate measurement gives way to some, and often considerable, guesswork as soon as the effect of the expenditure or disincentive is consolidated with other data or achieved through non-inclusion of information on the return. Even when data is available, the actual impact of tax expenditures and disincentives in terms of distributional burden across different categories of taxpayers, different income levels, etc, is not revealed in most jurisdictions’ tax expenditure budgets; only the estimated total cost is shown. To the extent that some extrapolation is necessary to estimate the overall fiscal cost of tax outcomes, the processes of costing both implicit and explicit tax expenditures poses difficulty.
As is the case with explicit statutory tax expenditures, the difficulty of costing the impact of a deviation from the benchmark in no way undermines the importance of the process of considering the possible rationales for deviation and asking whether the tax incentive or disincentive is the most cost-effective, economically efficient and fair method of achieving the possible objectives of the shift from a neutral tax. Identification of the implicit tax expenditure is a necessary first step to that analysis.
While implicit tax expenditures derive initially from judicial interpretations of terms in the tax law, some may be traced to subsequent assumptions by the tax administration as to how judicial doctrines will apply to transactions that have not been directly considered by the courts. The ATO explains its interpretation of the legislation by way of binding (on the Commissioner) rulings that have the force of law if they are more generous than the law intended. There may be a concern that inclusion in the tax expenditure statement produced by the Treasury of operational expenditures and disincentives resulting from ATO rulings will be perceived as implicit criticism by Treasury of the ATO for its failure to support the government’s tax policy. If the effect of an expanded tax expenditure analysis were to drive a wedge between the ATO and Treasury, the cost may outweigh any benefits from a shift in policy.
A perception that including the impact of ATO rulings in the tax expenditure account amounts to criticism of the ATO for undermining the policy of the tax law would be misconceived. In most cases, identification of operational shifts from the benchmark will not expose a failure by the tax administration to apply the law as it was intended to operate but rather a failure by the Treasury to take account of the likely judicial interpretation of the words it approved for inclusion in the law. The ATO is not responsible for writing the law, merely for applying it. It has no option but to adopt an understanding of the law most consistent with judicial precedents and approaches — the interpretation that would be expected from an appellate court. Once the ATO’s interpretation of the law is communicated through its public documents such as rulings and decisions, inaction by the government can and should be regarded as implicit endorsement of the deviation from the benchmark that has been identified by the ATO. At this point, it is the government that takes responsibility for the policy implications of the interpretation, not the ATO.
In a mass assessment tax regime subject to constant legislative change and judicial developments, there will, of course, inevitably be some instances when a decision by the tax administration is just plain wrong — the drafters of a ruling have simply misinterpreted the holding of a decision or the language of a statute. Tax expenditure analysis of these cases will show the administrative interpretations as errors, not revelations of implicit tax expenditures. In these instances, the concern that one agency is publicly criticising another would have merit. On the other hand, the prospect of inclusion of these things in the tax expenditure statement may be sufficient to minimise the problem. The risk of tax expenditure analysis of administrative decisions may prompt the ATO to revise its drafting and review processes and in the process reduce or even eliminate instances where the positions taken are difficult to justify.
Some implicit tax expenditures and disincentives will arise in areas in which the delineation of the neutral benchmark income tax is the subject of some debate. Where the benchmark has uncertain borders, the process of identifying and analysing expenditures and disincentives is inherently complicated. While similar difficulties can be observed with many explicit statutory tax expenditures, implicit tax expenditures may be more likely to touch upon issues sitting on the boundaries of the generally accepted benchmark principles. In both cases, however, the fact that questions may arise on the outer boundaries of a concept does not invalidate the use of the concept. Virtually every law on the books involves concepts that have uncertain boundaries at the extremes. The fact that the outer boundaries of concepts used to evaluate laws are uncertain does not prevent analysis of the vast majority of cases that fall within accepted benchmarks. The same is true of implicit tax expenditures that in most cases give rise to clear departures from generally accepted tax benchmark principles.
The introduction of explicit tax expenditures is inevitably accompanied by supplementary materials — parliamentary debates, second reading speeches, explanatory memoranda, etc — that provide plausible, if sometimes disingenuous, explanations for the measures. No similar guidance is available to assist analysis of implicit tax expenditures and if understanding the purpose of a tax expenditure is a central feature of tax expenditure analysis, the case for extending analysis to implicit tax expenditures would be weak.
The absence of official rationales for implicit tax expenditures and disincentives does not necessarily make the challenge of attributing an aim to the expenditures or disincentives any more difficult than is the case for many explicit tax expenditures, however. Quite often there is a muddle of alternative rationales offered for explicit tax concessions, reflecting the different views and ideologies of the range of players who collectively sponsor particular measures. The final rationale offered in an explanatory memorandum or second reading speech may bear little relation to the rationale offered by the proponent of a measure when the proposal is first raised. Currently tax expenditure budgets often deal with this issue by simply noting tax expenditures, with little, or more often no, credible rationale for the concessions.
The adoption of the current statutory capital gains concession for individuals provides a useful illustration of how difficult it may be to discover the purpose of an explicit tax expenditure. No rationale was offered for this concession when it was first proposed by the Australian government and none was included in the announcement establishing a business tax review to consider the concession or in the terms of reference for the Review of Business Taxation. The Review recommended a 50 per cent exemption for realised capital gains and finally offered a possible rationale for the concession: ‘to enliven and invigorate the Australian equities markets, to stimulate greater participation by individuals, and to achieve a better allocation of the nation’s capital resources.’ The second reading speech introducing a legislative amendment to achieve the concession offered a completely different rationale, the reduction of complexity with the end of cost base inflation adjustment. The Explanatory Memorandum accompanying the amendment Bill introducing the concession suggests the change was made to achieve ‘a simpler and more transparent system; how capital gains are worked out becomes easier to understand and apply’. The concession, one of the most significant recorded concessions, is presented without a rationale in the government’s annual tax expenditure budget.
In the first instance, it is sufficient for the purpose of the tax expenditure budget that the concession be identified as a deviation from the benchmark income tax that will affect taxpayer behaviour. To be sure, consideration of the possible objectives of a concession is an important second step. However, anticipated difficulties in understanding a rationale for the deviation should inhibit identification for neither implicit nor explicit tax expenditures. Indeed, consideration of possible rationales for tax expenditures can be fruitfully incorporated into subsequent stages of evaluating the operation of the concessions.
The neutral benchmark tax bases include features that could be difficult to apply in practice and as a result the tax legislation in most jurisdictions has adopted pragmatic operational benchmarks that accommodate perceived practical constraints on administering the neutral benchmark bases. Tax expenditure budgets may adopt the operational benchmark as the most practicable tax base from an administrative perspective while conceding it deviates from the truly neutral benchmark. Some judicial or administrative decisions may move the operation of the tax law away from the compromise operational base in favour of the conceptually preferable neutral benchmark base. These cases raise the question of whether deviations from the compromise base are themselves tax expenditures or instead should be viewed as corrective measures from a base that is in fact a tax expenditure.
The most significant operational departure from a benchmark tax system in Australia is the realisation basis used to recognise gains and losses on some investment assets. The neutral benchmark tax base is indifferent to portfolio choice and would treat the person who chooses to retain an appreciated or depreciated asset at the end of the tax period similarly to one who disposes of the asset during the year. That is, in the benchmark tax, accrued gains and losses enjoyed or suffered by the person who decides to retain assets are taxed in the same manner as realised gains and losses enjoyed or suffered by the person who changes investments. In practice, however, it is thought that valuation challenges would make it difficult or impossible to apply full accruals taxation to all assets. Accordingly, the operational base substitutes alternatives to the benchmark treatment.
For some types of appreciating and depreciating assets, the tax law prescribes surrogate measurements of annual gains and losses in lieu of actual annual valuation. The end of year value of wasting business assets, for example, is calculated using presumed depreciation formulas, while the annual change in value of debt instruments and various substitutes for debt is calculated by prorating the anticipated gain or loss on a compound basis over the life of the instrument. Gains and losses on trading stock and most investment assets apart from debt are most commonly taxed only on a realisation basis.
There are, as a result, two parallel bases used in the income tax law, a realisation base for some assets and an accrual base for others, with the accrual base using estimated values in some cases. This raises the question whether judicial or administrative decisions that recharacterise assets so they shift from one camp to another should be evaluated in terms of the operational base or the neutral benchmark base — a decision undermining the integrity of the operational base could reinforce the benchmark base, for example.
A variation of this question arises in cases where judicial doctrines have been codified into the legislation and later decisions move outcomes back towards the neutral benchmark. An example is the treatment in some cases of defence of title expenses. As noted earlier, these expenses were once treated as deductible revenue outgoings but a shift in judicial reasoning led to their characterisation as capital expenses and this outcome was codified with measures that require taxpayers to add these expenses to the cost base of assets if they can be tied to particular assets. A judicial decision characterising these costs as ordinary deductible business outgoings likely to arise as part of day-to-day business operations (even if they only arise once) would appear to undermine the operational benchmark but be in complete conformity with the benchmark base.
The problem also arises in the indirect tax base, most commonly in respect of financial supplies. While the goods and services tax (‘GST’) is intended to be a tax levied on final consumption, it is imposed at the point of supply on the supplier, not on the consumer. For most types of goods and services, there is a connection between the consideration received by the supplier and the value of the goods or services acquired by the consumer, so calculating GST on the basis of consideration paid captures the value of consumption.
The service provided by a financial institution is an exception to this rule. Financial institutions provide intermediary services, connecting individuals with funds to lend with borrowers who want to use those funds. The fee for the intermediary services provided by financial institutions is the spread between the interest rates they charge borrowers and the rates they pay to lenders. This type of service does not fit easily into the normal GST system of taxing supplies and providing input tax credits for business customers, mostly because the value of the service provided by the financial intermediary is different from the value of the payments to and from the service provider, the usual base for calculating a GST liability. Also, one group of persons involved in the transactions — private households with deposits in financial institutions — are not registered for GST purposes and there is no simple way to relieve them of tax on the services they enjoy even though those services relate to savings, not consumption. The pragmatic compromise developed first by the European Union and subsequently copied by almost all countries levying a GST, including Australia, is to adopt a system of ‘input taxation’ for financial intermediation services — that is, impose no GST on financial services but deny financial service providers full input tax credits for the tax borne on their acquisitions. The effect is that all persons acquiring financial services — depositors and borrowers, businesses and private individuals — pay some, but not full, GST on the services they use.
The normal GST system of taxing supplies and providing business tax customers with refundable credits achieves the benchmark goals of removing the tax from businesses along the production and sales chain and imposing it on final consumers. The operational benchmark used for financial intermediary services — subjecting the services to input taxation — overtaxes business consumers of the services and undertaxes final consumers. The distortion caused by the overtaxation of businesses can be serious as it reduces the competitiveness of enterprises relying on debt finance vis-a-vis those able to draw on equity and penalises those sectors that have more reliance on debt along the supply chain than do other sectors with different financing arrangements.
To address these issues, some countries, such as New Zealand and Singapore, have rules in place to effectively eliminate any taxation of business-to-business financial supplies, thus restoring the full intended neutrality of the GST. Australia has moved partly in this direction with limited credits for financial institutions, and the recent Henry Review recommended adoption of a system that would effectively provide full credits for GST borne by business customers of banks.
The norm in most countries and the basic default rule in Australia, however, is to input tax financial supplies, which raises the question of how decisions or measures that remove business-to-business transactions from the GST net should be characterised. An example is the Australian treatment of ‘reciprocal repurchase’ arrangements. A reciprocal repurchase arrangement, commonly called a ‘repo’ transaction in commercial terms, is a form of secured loan. Under a standard repo transaction, a borrower ‘sells’ an asset to the lender and at the same time provides a put option that will result in the borrower ‘repurchasing’ the asset after a specified period. The repurchase price includes a notional interest charge based on an appropriate commercial interest rate. While in form the repo transaction involves the sale of an asset to a lender and its repurchase at a higher price, in substance it is identical to a secured loan. GST legislation would thus normally treat these transactions as exempt financial supplies, and the Australian regulation setting out examples of exempt financial supplies specifically includes a reciprocal repurchase transaction as a type of financial supply.
Despite this, the ATO in an administrative decision has indicated it will allow a taxpayer entering into a repo arrangement that was not labelled as such in the ruling request to treat the transaction as a taxable sale and taxable repurchase. Quite likely the ATO officers responsible for the decision did not realise that the transaction in question fell within the scope of the regulation. On its face, the ATO interpretation looks to be a mistake — the law is explicit on the point and the decision is wrong in terms of the law, which establishes input taxation (that is, partial taxation) of financial supplies as the operational benchmark. As the parties to the transaction considered by the ATO were all registered businesses entitled to input tax credits for any tax imposed on their acquisitions, the effect of the ATO decision was to eliminate any tax burden for the de facto borrower.
But for what was probably an inadvertent error, the ATO would have sought to enforce the law and this decision could therefore be analysed as an implicit tax expenditure, extending a concession where none was intended under the operational benchmark or under ordinary ATO practice. However, the transaction in question is only likely to be undertaken where there are GST-registered businesses on both sides of the transaction and the outcome — allowing both the borrower and lender to recover fully all input taxes related to the transaction, thus making it a tax-free business-to-business transaction — is entirely consistent with the benchmark principles of a GST, namely that the tax should only stick to supplies to final consumers. In other words, failure to implement the law as set out achieves the policy objectives of the model GST rather than the operational benchmark. In this case, the apparent mistake could be seen as a preferable outcome. In effect, it allows business taxpayers a route to opt out of the ordinary GST rules and achieve a desirable tax-free outcome.
It is unclear whether overall the outcome is positive or negative. This is not, however, a reason for concluding there is no merit to tax expenditure analysis of the decision. To the contrary, the very fact that the uncertainty in analysis exists is itself of great value. Tax expenditure analysis will expose for critical scrutiny the difference between the operational benchmark and the neutral benchmark and show that taxpayers have, with the inadvertent assistance of the tax authority, found a simple way to remove the GST from business-to-business financial supplies. In this case, the analysis may trigger a rethink of the operational benchmark and prompt the government to explore alternative structures that can shift the tax towards the neutral benchmark, as has occurred in some other jurisdictions. More generally, the example shows that there are inherent benefits of exposure when the operational benchmark deviates from the model benchmark. The fact that a decision shifts outcomes from an operational benchmark to a model benchmark renders the decision a prime candidate for critical examination.
Outsiders sometimes wonder if tax law in practice is little more than a jumble of apparently inconsistent rules featuring innumerable overlaps and lacunae. Insiders know this to be the case. In the case of explicit deviations from a benchmark tax, the legislature devotes much time inserting non-revenue measures into tax laws and then following up with continuous amendment to slow the haemorrhage as well-advised taxpayers aggressively test the ill-defined borders of concessions. In the case of implicit deviations resulting from judicial decisions, the legislature’s response is only reactive, patching the loopholes as taxpayers restructure arrangements to take advantage of the judiciary’s generosity.
Identification and critical analysis of implicit tax concessions in the government’s annual tax expenditure statement could set the stage for meaningful reform of much of the law. The benefits would be limited, however, if this analysis was treated as an end in itself or merely triggered another round of limited responses to address the most egregious problems identified. One of the prime causes of complexity and uncertainty in Australian taxation is the proliferation of piecemeal responses that bypass a serious examination of the underlying structural issues. Tax expenditure analysis of implicit tax expenditures could end up reinforcing the problem if it led to measures that address the symptoms rather than prompt reform of the structural shortcomings.
The risk of doing more harm than good with an extension of tax expenditure analysis can be illustrated by considering its potential application to one area of tax law described earlier, the treatment of capital expenses. The basic statutory rule prevents taxpayers from immediately deducting capital expenses, with other provisions providing for the later recognition of these outlays. Spotting the concessions among the statutory amortisation rules for capital expenses is not difficult. In the benchmark income tax, expenses to acquire wasting assets are recognised over the life of the benefits acquired. A number of Australian statutory measures are consistent with the benchmark: borrowing expenses are amortised over the life of the loan to which they relate; the acquisition cost of intellectual property is deductible over the legal life of the property; and generally the cost of acquiring tangible equipment is deductible over the estimated usable life of the equipment. Fixed amortisation periods far shorter than actual life are specified for assets such as tractors (6 years and 8 months) and telephone lines or powerlines connected to a farm (10 years), while lease expenses are fully deductible at the start of the lease.
The short amortisation times for tractors and utility connections and the immediate write-off for lease expenses are unambiguously concessional rules. Tax expenditure analysis of rules that provide accelerated recognition of capital expenses is relatively straightforward. Amortisation of the cost of installing powerlines over 10 years when the known life of the benefit is much longer can be treated as a concession to foster rural development, as can the rapid write-off for tractors. Up-front deduction for lease expenses can be analysed as a somewhat perverse subsidy for persons shifting lease costs to the commencement of the lease, and so on.
The benchmark income tax is also the starting point for identifying implicit expenditures and disincentives. For example, in the benchmark income tax, a lump sum payment to acquire a multi-year tied-house agreement would be recognised over the life of the contract. Under judicial doctrines, the cost of a relatively short-term agreement is characterised as an immediately deductible revenue expense while prepayment for a longer multi-year agreement is characterised as a capital expense. The statutory response to the first characterisation is a rule denying up-front deductions for the acquisition costs and instead mandating amortisation over the life of the contract. The statutory response to the second characterisation allows a taxpayer incurring an expense of this sort to recognise the cost of acquiring a longer-life agreement at the expiry of the agreement by way of a capital loss equal to the cost of the acquisition. The first response is consistent with the benchmark treatment and gives rise to no tax expenditure or tax disincentive. A tax expenditure analysis would cast the second rule as a tax disincentive for persons paying lump sums to obtain commercially appropriate wasting benefits.
Similarly, in the benchmark income tax the cost of litigation to defend an attack on title to an asset is a currently deductible expense as it yields no asset or benefit lasting into the future. Tax expenditure analysis would regard a rule forcing taxpayers to capitalise into the price of assets expenses incurred to protect legal title as a (highly irrational) tax disincentive to penalise enterprises that respond to threats to their title to assets.
Relatively simple proposals for reform would follow if tax expenditure analysis were applied to implicit tax expenditures in the same manner as it is used for explicit tax expenditures. One recommendation would be to abolish the rule requiring taxpayers to capitalise the cost of defending title and restore the deductibility of these outgoings. Another would be to replace the rule requiring capitalisation of the cost of longer-term contracts with a measure extending the amortisation rule for prepayments of shorter-term contracts to include prepayments enjoying a capital characterisation under judicial tests.
Reforms of this type would disregard a fundamental difference between explicit and implicit tax expenditures and disincentives, however. Explicit tax expenditures derive from deliberate decisions to deviate from a benchmark income tax to favour particular taxpayers or particular transactions. These are akin to distinct expenditure or disincentive programs and their effectiveness or fairness can be improved directly through better targeted rules. In contrast, modifying statutory rules enacted in response to incorporation of judicial concepts into the tax system ignores completely the underlying issue that prompted the questionable rules in the first place — the judicial doctrines, developed in the context of an inadequate statutory framework, that characterise receipts and outgoings using considerations unrelated to benchmark tax design principles. Instead of tax expenditure analysis being used to develop piecemeal modifications to inappropriate rules, it can be used as a tool to unearth and address the underlying problems.
An examination of the judicial doctrines that gave rise to statutory responses and the ad hoc responses reveals two underlying problems in this area of tax law. First, the analysis shows that the judicial tests to distinguish capital and revenue expenses yield inappropriate results in many cases. Second, the analysis reveals an incomplete amortisation regime with significant gaps and no overriding principle, establishing separate rules for different types of wasting assets. The solution to the first problem is not to enact specific rules to address every type of mischaracterised expense separately but instead to replace the judicial test with a statutory definition of capital expenses based on the longevity of the benefits acquired. The solution to the second problem is to replace multiple amortisation rules with a single principle: the cost of a wasting asset or benefit acquired to derive assessable income is recognised over the life of the asset.
A tax law built upon these foundations can accommodate any explicit or implicit tax expenditures or disincentives desired. If the legislature wants to subsidise farmers hooking up to mains electricity, a 10 year write-off can be set out as an explicit deviation from the rule. If it wanted to penalise those who take the initiative to protect their title to their assets, it can similarly deny an immediate write-off for the expense. But in each case, the legislature would have to explicitly carve out the concession or penalty from the general rule and, hopefully, provide a rationale that can help target the exception. Most importantly, the general rule will be transparent and known to all who do not fall into one of the deliberate carve-outs.
Proposals to extend the scope of tax expenditure analysis and to use this analysis to support a shift towards principle-based design of tax law will almost certainly encounter considerable resistance from a wide coalition of groups with complementary interests. Politicians would have to confront directly issues they prefer to sweep aside to be dealt with by administrators. Courts handing down decisions and administrators issuing rulings that give rise to implicit tax expenditures will be confronted with timely critical analysis of their judgments and rulings respectively, an outcome to which the former group may be indifferent and that the latter group could find distressing. To the extent that problems may be traced to bad drafting instructions, the department responsible for poor advice would be the same one documenting the consequences of the design shortcomings and use of sub-optimal language. Finally, the extension of expenditure analysis to include implicit tax expenditures deriving from administrative and judicial decisions and legislative measures that adopt judicial characterisations will have significant resource implications for Treasury.
As difficult as it may be to extend the current system of tax expenditure analysis to include implicit tax expenditure, the long-term benefits could greatly outweigh the costs if exposure of implicit concessions prompts reform to remove or better target the reliefs. The initial revenue costs of inadvertent and untargeted subsidy programs may pale into insignificance when put beside the deadweight costs of planning and restructuring transactions and investments into preferentially taxed regimes. These costs are in turn quite probably overshadowed by the costs to society of tax-driven distortions as human and financial capital are inefficiently reallocated to exploit the implicit concessions. A program that catalogues and analyses the distortions will not in itself fix the problems but it is a necessary first step to genuine reform.
The most important gain from establishing a process to analyse implicit tax expenditures may not be the direct benefits realised through reform of the concessions and disincentives but rather may arise from the impact this process could have on tax law drafting, administration and adjudication. Preparation of an expanded tax expenditure budget will prompt changes to the interpretation of current law by all parties and help facilitate a shift to principle-based tax law design. The sooner a shift is made to full tax expenditure analysis, the sooner tax law, administration and adjudication will improve.
[*] LLB (Osgoode), LLM (Harv); Professor of Taxation, Monash University; Director, Monash University Taxation Law and Policy Research Institute.
 The expression ‘tax expenditure budget’ is commonly attributed to Surrey’s first academic paper on the topic (see Stanley S Surrey, ‘Tax Incentives as a Device for Implementing Government Policy: A Comparison with Direct Government Expenditures’ (1970) 83 Harvard Law Review 705), although its origins are sometimes attributed to earlier German work. The phrase was used two years prior to Surrey’s Harvard Law Review paper by the United States Treasury in Treasury Department, United States, Annual Report of the Secretary of the Treasury on the State of the Finances for the Fiscal Year Ended June 30, 1968 (Document No 3245, 1969) 326–40. See also Stanley S Surrey, Pathways to Tax Reform: The Concept of Tax Expenditures (Harvard University Press, 1973), which played an important role in extending the understanding of the concept. The concept has been the subject of an enormous volume of literature in the past four decades. Much of this is covered in J Clifton Fleming Jr and Robert J Peroni, ‘Reinvigorating Tax Expenditure Analysis and Its International Dimension’ (2008) 27 Virginia Tax Review 437.
 See Surrey, ‘Tax Incentives’, above n 1, 706; Mark Burton, ‘Making the Australian Tax Expenditures Statement an Effective Policy Instrument — From Fiscal Record to Transparent Report’  JlATax 1; (2005) 8 Journal of Australian Taxation 1, 1–3.
 The Australian experience is set out in Burton, above n 2; Kerrie Sadiq, ‘The Implementation of Social and Economic Policy through the Tax Regime: A Review of Australia’s Tax Expenditures Program’ (2008) 23 Australian Tax Forum 339. There is, to be sure, a vibrant debate in academic literature about the definition of a ‘benchmark’ tax base. An early piece by Victor Thuronyi summarised much of this debate nicely: see Victor Thuronyi, ‘Tax Expenditures: A Reassessment’  Duke Law Journal 1155. A key component in the debate is whether capital gains should be measured on an accrual or a realisation basis. In practical terms, however, the debate concerns issues that are truly at the margin — there is little or no debate about the classification of the vast majority of tax expenditures identified in national tax expenditure accounts.
 Examples include several measures seeking to impose moral limitations on the way taxpayers conduct business by denying deductions for business expenses in the nature of fines or bribes: see Income Tax Assessment Act 1997 (Cth) ss 26-5 (fines), 26-52 (bribes to foreign public officials), 26-53 (bribes to public officials).
 See Surrey, Pathways to Tax Reform, above n 1, 336 n 61.
 Richard and Peggy Musgrave, in their seminal work on public finance, describe this principle as follows: ‘Taxes should be chosen so as to minimize interference with economic decisions in otherwise efficient markets’ (Richard A Musgrave and Peggy B Musgrave, Public Finance in Theory and Practice (McGraw-Hill, 2nd ed, 1976) 210).
 See Sadiq, above n 3, 340–1; Surrey, Pathways to Tax Reform, above n 1, 6.
 The relationship between tax concessions and tax law complexity was explored in Tracy Oliver and Scott Bartley, ‘Tax System Complexity and Compliance Costs — Some Theoretical Considerations’  (Winter) Economic Roundup 53.
 For discussion in the context of Australia, see Sadiq, above n 3.
 Deductions and exemptions are inherently more valuable to high income persons than to low income persons.
 For example, personal deductions for primary earners supporting spouses or child housekeepers have been replaced with a disappearing rebate that shades out as the claimant’s income rises: see Income Tax Assessment Act 1936 (Cth) s 159J. The rebate is slowly being phased out altogether. Concessional deductions to support the production of Australian films have been replaced with refundable tax credits (called ‘offsets’) in Income Tax Assessment Act 1997 (Cth) div 376 and a direct grant program administered by a dedicated Commonwealth agency, Screen Australia, with further funding provided by state Screen agencies.
 A recent review of complexity in Australia’s tax system may be found in Chris Evans and Binh Tran-Nam, ‘Controlling Tax Complexity: Rhetoric or Reality?’ in Chris Evans, Richard Krever and Peter Mellor (eds), Australia’s Future Tax System: The Prospects after Henry (Thomson Reuters, 2010) 439, 443–4. A leading conceptual study of complexity in tax law remains Graeme S Cooper, ‘Themes and Issues in Tax Simplification’ (1993) 10 Australian Tax Forum 417. An analysis of causes of complexity may be found in Richard Krever, ‘Taming Complexity in Australian Income Tax’  SydLawRw 22; (2003) 25 Sydney Law Review 467, 469–90.
 Surrey, ‘Tax Incentives’, above n 1, 706; Surrey, Pathways to Tax Reform, above n 1, 3–5, 35–9.
 Surrey, ‘Tax Incentives’, above n 1, 706; Surrey, Pathways to Tax Reform, above n 1, 36.
 Surrey, ‘Tax Incentives’, above n 1, 706; Surrey, Pathways to Tax Reform, above n 1, 100–2.
 Surrey, ‘Tax Incentives’, above n 1, 706; Surrey, Pathways to Tax Reform, above n 1, 6. Surrey also included preferential tax rates in his classification: Surrey, ‘Tax Incentives’, above n 1, 706.
 Internal Revenue Code of 1986, 26 USC § 63(a) (taxable income means gross income minus deductions); Income Tax Assessment Act 1997 (Cth) s 4-15 (subtract your deductions from your assessable income and the result is your taxable income).
 Internal Revenue Code of 1986, 26 USC § 263(a); Income Tax Assessment Act 1997 (Cth) s 8-1(2)(a).
 Internal Revenue Code of 1986, 26 USC §§ 161 (immediate deduction), 167(a) (recognition over time) (2006); Income Tax Assessment Act 1997 (Cth) s 8-1(1) (immediate deduction), pt 2-10 (recognition over time).
 The adoption of a tax expenditure budget in Australia followed the report of the House of Representatives Standing Committee on Expenditure, Parliament of Australia, Taxation Expenditures (1982), tabled in Commonwealth Parliament on 16 September 1982. Treasury trod carefully, first adding an appendix to the budget papers describing tax expenditures and finally after two years issuing a separate tax expenditure statement.
 Internal Revenue Code of 1986, 26 USC §63(a) (2006).
 Commissioner of Internal Revenue v Glenshaw Glass Co USSC 43; , 348 US 426, 429–30 (Warren CJ for Warren CJ, Black, Reed, Frankfurter, Douglas, Burton, Clark, Minton and Harlan JJ) (1955).
 See, eg, Cesarini v United States USCA6 292; , 428 F 2d 812 (Phillips CJ, McCree and O’Sullivan JJ) (6th Cir, 1970).
 Gifts fall within the US judicial concept of income but are explicitly taken back out by the US Internal Revenue Code of 1986, 26 USC § 102(a) (2006) to avoid double taxation, as the payments are potentially subject to the gift tax.
 See ibid sub-tit A ch 1 sub-ch B pt III.
 In the case of the capital gains concession, the tax expenditure is identified by the lower tax rate applied to capital gains: ibid §1(h) provides a 15 per cent rate for most capital gains for most individuals, while the highest marginal rate on other income under § 1 is 35 per cent.
 The term derives from and the concept is best articulated in the leading article by Neil Brooks, ‘The Responsibility of Judges in Interpreting Tax Legislation’ in Graeme S Cooper (ed), Tax Avoidance and the Rule of Law (IBFD, 1997) 93, 122–4.
 John F Avery Jones et al, ‘Treaty Conflicts in Categorizing Income as Business Profits Caused by Differences in Approach between Common Law and Civil Law’ (2003) 57 Bulletin for International Fiscal Documentation 237, 237 n 3; Ross W Parsons, ‘Income Taxation — An Institution in Decay?’ (1986) 3 Australian Tax Forum 233, reproduced in ‘Income Taxation — An Institution in Decay?’  MonashULawRw 6; (1986) 12 Monash University Law Review 77 and later in a slightly revised version in ‘Income Taxation: An Institution in Decay’  SydLawRw 28; (1991) 13 Sydney Law Review 435. A New Zealand scholar, John Prebble, has argued Parsons was incorrect in attributing the judicial concept of income to trust law notions, asserting instead that the judicial notion is simply the ‘ordinary’ meaning of income, although he concedes the income concept accepted for tax purposes is the same as that used for trust purposes: see John Prebble, ‘Income Taxation: A Structure Built on Sand’ (2002) 24 Sydney Law Review 301, 302–4.
 See generally J D Heydon and P L Loughlan, Equity and Trusts: Cases and Materials (LexisNexis Butterworths, 7th ed, 2007) ch 33.
 This process is described in Richard Krever, ‘Interpreting Income Tax Laws in the Common Law World’ in Markus Achatz et al (eds), Steuerrecht Verfassungsrecht Europarecht (Facultas Verlags- und Buchhandels, 2007) 354, 357–63.
 See ibid 369; Rick Krever, ‘The Ironic Australian Legacy of Eisner v Macomber’ (1990) 7 Australian Tax Forum 191, 195.
 Krever, ‘The Ironic Australian Legacy of Eisner v Macomber’, above n 31, 193.
 See, eg, the first general income tax statute in the Australasian colonies, South Australia’s Taxation Act 1884 (SA), with a global structure that was carried to the income tax when it was enacted at the federal level in Australia in the Income Tax Assessment Act 1915 (Cth): Cynthia Coleman and Margaret McKerchar, ‘The Chicken or the Egg? A Historical Review of the Influence of Tax Administration on the Development of Income Tax Law in Australia’ in John Tiley (ed), Studies in the History of Tax Law (Hart Publishing, 2004) 285, 287–8. The rejection of the UK model is reviewed in Peter Mellor, ‘Origins of the Judicial Concept of Income in Australia’ (2010) 25 Australian Tax Forum 339, 341–2. One motivation for rejection of the UK schedular model may have been commitment of Australian politicians to the goal of progressivity based on progressive rates applying to a taxpayer’s total income. See J A L Gunn, Commonwealth Income Tax Law and Practice (Butterworth & Co, 1943) 2–3.
 A classic example is the assumption in Australia that the non-competition payment received by the taxpayer in Higgs (Inspector of Taxes) v Olivier  1 Ch 311, which was found to be outside the scope of Income Tax Act 1918, 8 & 9 Geo 5, c 40, sch D case II, would fall outside the Australian judicial concept of income. The presumption that Australian judicial concepts automatically followed decisions of UK judges considering the very different UK statute was once questioned by Kitto J in Dickenson v Federal Commissioner of Taxation  HCA 62; (1958) 98 CLR 460, 492, but the capital gains character of non-competition payments has been accepted by the Australian Treasury, which deliberately excluded these payments from the definition of assessable eligible termination payments and assessable fringe benefits, leaving them to be taxed in the capital gains provisions of the Income Tax Assessment Act 1997 (Cth): see Income Tax Assessment Act 1997 (Cth) s 82-135(j); Fringe Benefits Tax Assessment Act 1986 (Cth) s 136(1) (definition of ‘fringe benefit’, para (m)).
 See Krever, ‘Interpreting Income Tax Laws’, above n 30, 369–70.
 Income Tax Assessment Act 1997 (Cth) s 8-1(2)(a), pt 2-10; Internal Revenue Code of 1986, 26 USC §§ 263, 167(a) (2006).
 The decision of the Seventh Circuit Court of Appeals in Encyclopaedia Britannica Inc v Commissioner of Internal Revenue USCA7 590; , 685 F 2d 212 (7th Cir, 1982) explains the logic behind this approach clearly. The Court found the taxpayer’s expenses to acquire a book manuscript a capital expense because the acquisition ‘was intended to yield Encyclopaedia Britannica income over a period of years. … Where the income is generated over a period of years the expenditures should be classified as capital’: at 214 (Posner J).
 The most cited articulation of the principles of characterisation of expenses is that of Dixon J in Sun Newspapers Ltd v Federal Commissioner of Taxation  HCA 73; (1938) 61 CLR 337. Surveying the precedents, Dixon J found three matters were important to the revenue or capital nature of expenses: the character of the advantage sought as a consequence of an outlay, the manner in which the benefit would be used, and whether it was acquired by way of a lump sum payment or periodic payments: at 363. The longevity of the benefit acquired was merely a factor that might be relevant to the first of the three relevant matters.
 The development of Australian tests is reviewed in Philip Burgess et al, Income Taxation — Commentary and Materials (Lawbook, 6th ed, 2009) 493–500.
 In tax policy terms, there is no deduction for the cost of acquiring an asset with a life past the end of the tax year, because there has been no actual decrease in economic wealth; the taxpayer has merely substituted one asset (cash) for another (the tangible or intangible benefit with a value past the end of the year). A neutral income tax measures expenditures as economic capacity is actually consumed: a machine would be depreciated as it lost value over its effective life; an intangible asset would be amortised as it lost value due to the passing of time as it nears the end of its legal life; and the cost of a non-wasting asset such as land would be recognised when there is a disposal of the asset.
 A challenge to title will reduce the value of the taxpayer’s asset. If the taxpayer loses the challenge, the value is permanently lost and there clearly is no ongoing benefit from the expenditure. If the taxpayer wins, the asset’s value is restored to the value it had prior to the challenge. Accounting principles assume there is no ongoing value, however, as meeting the challenge of one attack on legal title brings no new asset or benefit into existence. For a review of the common position in jurisdictions outside Australia and the shift in Australia from this view to the current view that these expenses have a ‘capital’ nature, see Richard Krever, ‘Capital or Current: The Tax Treatment of Expenditures to Preserve a Taxpayer’s Title or Interest in Assets’  MonashULawRw 5; (1986) 12 Monash University Law Review 49.
 While the test advocated by Dixon CJ, attributing a capital or revenue character to an expense depending on whether it seemed to be associated with the taxpayer’s business ‘process’ or its business ‘structure’, was not new, his clear articulation paved the way for its eventual triumph. Dixon J first argued for this approach to trump all other considerations in a concurring decision in Sun Newspapers Ltd v Federal Commissioner of Taxation  HCA 73; (1938) 61 CLR 337, 359–63. It was not accepted by a majority of the High Court in Hallstroms Pty Ltd v Federal Commissioner of Taxation  HCA 34; (1946) 72 CLR 634, 641 (Latham CJ), 643–4 (Starke J), 655 (Williams J) (the majority finding defence of title expenses to be revenue outgoings).
 Broken Hill Theatres Pty Ltd v Federal Commissioner of Taxation  HCA 75; (1952) 85 CLR 423, 433–4 (Dixon CJ, McTiernan, Fullagar and Kitto JJ). The expenditure incurred by the taxpayer in Broken Hill Theatres continues to be characterised as capital outgoings, though today it can be amortised over five years under Income Tax Assessment Act 1997 (Cth) s 40-880.
 Tax Law Improvement Act (No 1) 1998 (Cth) sch 1 item 1, inserting Income Tax Assessment Act 1997 (Cth) s 110-25(6).
 Income Tax Assessment Act 1997 (Cth) s 110-25(6).
 Higgs (Inspector of Taxes) v Olivier  1 Ch 311, 320 (Evershed MR).
 Kitto J of the High Court of Australia did suggest in obiter that the UK precedent might not be relevant in the face of the Australian legislation: Dickenson v Federal Commissioner of Taxation  HCA 62; (1958) 98 CLR 460, 492; however, this comment had no lasting impact in Australia.
 See Fringe Benefits Tax Assessment Act 1986 (Cth) s 136(1) (definition of ‘fringe benefit’, para (m)).
 Income Tax Assessment Act 1936 (Cth) s 27A(1) (definition of ‘eligible termination payment’, sub-para (m)), as repealed by Superannuation Legislation Amendment (Simplification) Act 2007 (Cth) sch 1 item 3. See also the corresponding exclusion in Income Tax Assessment Act 1997 (Cth) s 82-135(j).
 Income Tax Assessment Act 1936 (Cth) s 160M(7), replaced by Income Tax Assessment Act 1997 (Cth) s 104-155, was originally intended to apply to situations involving the creation of rights in favour of other persons (treated as the disposal of an asset); Income Tax Assessment Act 1936 (Cth) s 160M(6) also applied more generally to disposals of assets that did not exist prior to the disposal.
 Income Tax Assessment Act 1997 (Cth) s 104-35.
 Ibid s 115-25(3)(a).
 Income Tax Assessment Act 1936 (Cth) s 54, repealed by Tax Laws Amendment (Repeal of Inoperative Provisions) Act 2006 (Cth) sch 1.
 Gunn, above n 33, 407.
 See especially the comments of Reid LJ in Regent Oil Co Ltd v Strick (Inspector of Taxes)  AC 295, 324.
 See, eg, BP Australia Ltd v Federal Commissioner of Taxation  UKPCHCA 2; (1965) 112 CLR 386 (Privy Council) (expenses incurred by petrol company to secure three to five year exclusive supplier contracts with retailers characterised as revenue outgoings).
 See, eg, Regent Oil Co Ltd v Strick (Inspector of Taxes)  AC 295 (expenses incurred by petrol company to secure 5, 10 and 21 year tied-house agreements by way of lease and lease-back contracts requiring only nominal rent from the retailer so long as the company’s petrol was sold on an exclusive basis characterised as capital outgoings).
 Income Tax Assessment Act 1936 (Cth) s 82KZMD.
 Ibid s 82KZL (definition of ‘excluded expenditure’, para (d)) excludes capital outgoings from the operation of the amortisation regime.
 If the cost of acquiring the asset (the multi-year contract) is characterised as a capital outgoing, the expense is treated as the reduced cost base of the asset: Income Tax Assessment Act 1997 (Cth) ss 110-25(2), 110-55(2). The expiry of the contract gives rise to a ‘CGT event’
(s 104-25(1)(c)), yielding a capital loss equal to the reduced cost base (s 104-25(3)).
 See The Treasury (Cth), Tax Expenditures Statement 2009 (2010) 4.
 See the discussion of tax administration in Burton, above n 2, 56–61.
 See Neil Brooks, ‘The Under-Appreciated Implications of the Tax Expenditure Concept’ in Chris Evans and Richard Krever (eds), Australian Business Tax Reform in Retrospect and Prospect (Thomson Reuters, 2009) 233, 234–7. Contributions to debate over this issue are surveyed in Fleming and Peroni, above n 1.
 While courts sometimes refer to explicit policy criteria in their decisions interpreting tax statutes, interpretation based on transplanted categories or outcomes are explained by reference to precedents, not published policy objectives.
 See Peter Costello, Treasurer (Cth), Tax Reform: Not a New Tax, A New Tax System — The Howard Government’s Plan for a New Tax System (1998), with a section entitled ‘Gains for Businesses’ proposing ‘[c]onsultations with the business sector on … the prospect of further CGT relief’ without any explanation for the proposal: at 25.
 Peter Costello, ‘Business Income Tax Consultation’ (Media Release, No 081, 14 August 1998), establishing a business tax review to, inter alia, examine the scope for ‘capping the rate of tax applying to capital gains for individuals at 30 per cent’, again, without any rationale for the proposed concession.
 See Review of Business Taxation, Report — A Tax System Redesigned: More Certain, Equitable and Durable (1999) vii.
 Ibid 598. See also at 599, noting that the 50 per cent exclusion would ‘reduce the effective top marginal rate on capital gains income to 24.25 per cent.’
 Commonwealth, Parliamentary Debates, House of Representatives, 21 October 1999, 12181 (Peter Costello, Treasurer).
 Explanatory Memorandum, New Business Tax System (Integrity and Other Measures) Bill 1999 (Cth) 148 [12.5].
 In the most recent financial year, the capital gains discount was estimated to be the equivalent of an indirect expenditure of $4.7 billion for investments in assets other than principal residences and a further $20 billion for investments in family homes. See The Treasury (Cth), Tax Expenditures Statement 2011 (2012) 7.
 The realisation basis for capital gains follows from the structure of the legislation which measures gains and losses on the transfer of an asset. The most important manifestation of the realisation basis is found in Income Tax Assessment Act 1997 (Cth) s 104-10, the primary capital gains recognition measure, which records gains and losses on the disposal of capital assets. There is a limited exception to this rule for some debt assets, for which gains are recognised on an accrual basis under Income Tax Assessment Act 1936 (Cth) pt III div 16E and Income Tax Assessment Act 1997 (Cth) div 230.
 Income Tax Assessment Act 1997 (Cth) div 40.
 Income Tax Assessment Act 1936 (Cth) pt III div 16E; ibid div 230.
 Taxpayers do have the option of recognising gains on trading stock on an accrual basis: see Income Tax Assessment Act 1997 (Cth) s 70-45. However, the option is rarely exercised unless the taxpayer anticipates a rise in the tax in a future year.
 See above nn 41–46 and accompanying text.
 See, eg, Federal Commissioner of Taxation v Consolidated Fertilizers Ltd (1991) 101 ALR 385.
 A New Tax System (Goods and Services Tax) Act 1999 (Cth) div 9.
 GST is payable on the value of a taxable supply (see ibid s 9-70) and the value is defined by reference to the gross consideration paid (see ibid s 9-75).
 See ibid pt 2-5.
 Ibid s 40-5; for discussion of the European approach, see Adrian Ogley, Principles of Value Added Tax: A European Perspective (Interfisc Publishing, 1998) 55–9.
 Details of the New Zealand rules are set out in Marie Pallot, ‘GST and Financial Services — Rating Zero-Rating’ in Richard Krever and David White (eds), GST in Retrospect and Prospect (Brookers, 2007) 163 and Marie Pallot, ‘GST and Financial Services in New Zealand’ in Christine Peacock (ed), GST in Australia: Looking Forward from the First Decade (Thomson Reuters, 2011) 161. The Singapore regime is described briefly in Satya Poddar, ‘VAT on Financial Services — Searching for a Workable Compromise’ in Richard Krever and David White (eds), GST in Retrospect and Prospect (Brookers, 2007) 179, 187–8.
 A New Tax System (Goods and Services Tax) Act 1999 (Cth) div 70.
 See Review Panel, Australia’s Future Tax System: Report to the Treasurer — Part Two: Detailed Analysis (2009) 303–13 (‘D4: Taxing Financial Services’).
 Stewart Karlinsky and Richard Krever, ‘Characterising Derivative-Based Loan Arrangements’ (2004) 19 Australian Tax Forum 435, 436–7.
 A New Tax System (Goods and Services Tax) Regulations 1999 (Cth) sch 1 pt 8 item 3.
 ATO, Goods and Services Tax: GST and Agreement for the Supply and Repurchase of a Commodity, ATO ID 2004/76, 12 August 2002 distinguishes a sale and repurchase of commodities from a sale and repurchase of securities on the basis of a definition of a reciprocal repurchase agreement in an earlier ATO Ruling (ATO, Goods and Services Tax: GST Treatment of Financial Supplies and Related Supplies and Acquisitions, GSTR 2002/2, 26 June 2002, sch 1 (definition of ‘Repurchase Agreement (Repos)’) rather than the commercial meaning consistent with the intent of the legislation. There is no indication in the ATO Ruling that the definition relied upon by the author of the ATO ID was intended to be an exclusive definition; rather, the authors of the ATO Ruling based their definition on the transactions in effect at that time.
 See above n 83.
 Income Tax Assessment Act 1997 (Cth) s 8-1(2)(a).
 Ibid pt 2-10.
 Ibid s 25-25.
 Ibid s 40-75(6).
 Ibid s 40-25.
 Ibid s 40-102.
 Ibid s 40-645.
 Ibid s 25-20.
 The most recent tax expenditure statement records the equivalent of a $17 million annual subsidy for this concession: see The Treasury (Cth), Tax Expenditure Statement 2010 (2011) 102.
 Income Tax Assessment Act 1997 (Cth) s 25-20.
 See above nn 57–58 and accompanying text.
 See above n 59.
 See above n 61.
 See above n 41.
 This principle-based approach to statutory design is used in the US tax law: see Internal Revenue Code of 1986, 26 USC §167 (2006). Greg Pinder has advocated its use in Australia in his analysis for Australian Treasury on principle-based law design: see Greg Pinder, ‘The Coherent Principles Approach to Tax Law Design’  (Autumn) Economic Roundup 75. The report of the Ralph Review of business taxation released in 1999 was accompanied by a proposal for a principle-based assessment Act: see Review of Business Taxation, A Tax System Redesigned: More Certain, Equitable and Durable — Draft Legislation (1999).