Corps de l’article

Introduction

Even though most Canadians may not think of taxes as optional, there are over 230 provisions in Canada’s Income Tax Act (the Act)[1] that give taxpayers a choice regarding how to calculate their taxes.[2] These “tax elections” are so ubiquitous as to sometimes go unnoticed as a matter of design and have generated little discussion among academics and tax experts in Canada.[3] However, ignoring the design of tax elections is perilous for a tax system that seeks to be both equitable (to engineer an after-tax distribution of resources that accords with social norms of fairness) and efficient (to raise a required amount of revenue at minimum cost).

I demonstrate in this paper that improperly structured tax elections in the Act can impose substantial transaction costs on taxpayers and the government, while squandering their intended benefits and disproportionately burdening those taxpayers least able to navigate the complexity of the election. From the standpoint of the taxpayer, a tax election offers flexibility and the potential to decrease her tax liability, but also requires her to invest resources in understanding the election. She must determine which available tax treatment is best, as well as learn how to comply with the election and bear the costs of filing it with the government (either by herself or through her tax representative).

From the standpoint of the government, another set of costs and benefits of tax elections is at play. On the cost side, tax elections require the government to invest resources in administering the election. The government must bear the costs of processing the elections that are filed by taxpayers, implementing the elective treatment as a basis for calculating taxes owed, and fielding requests from taxpayers who want to modify or withdraw tax elections (in cases where such changes to elective choices are permitted by law).

On the benefit side for the government, however, is the notion that elections can act as “screens”. In the tax context, a screen is a choice offered by government that forces (at least some) taxpayers to separate into two different groups.[4] The government’s resulting observation—into which group a taxpayer self-selects when presented with the tax choice—may prove useful in administering, enforcing, or tailoring the tax law.[5] In this paper, I will refer to such observations about the self-selection of taxpayers as “information”—defined narrowly as taxpayer-specific inferences that can be made by the government upon observing taxpayers’ elective choices. In cases where such information is produced by an election, the information’s measurable benefits must be traded off against the costs of having an election, whether or not it is structured as a penalty default—that is, as a default treatment that is contrary to what a majority of taxpayers would prefer.

Here, I analyze the trade-offs at the heart of tax elections to tackle the most basic—but arguably the most important—aspect of designing a tax election: how to determine the tax treatment that will be available to taxpayers by default. To distill a clear set of principles for policymakers, I examine the nature of the transaction costs presented by tax elections as well as the literature on default rules in the law, with an eye toward designing tax elections in a “least-cost” manner.[6] Specifically, this approach seeks to maximize the measurable benefits of a given tax election while minimizing the election’s measurable costs, such as transaction costs generally or complexity costs more specifically.[7] Such an approach, while limited in that it focuses only on measurable benefits and costs, has the benefit of allowing these to trade off against one another and be calculated on a net basis, thus ensuring that fewer measurable resources in society will be devoted to activities related to tax compliance, and instead resources can be put to more productive uses.

My examination results in a straightforward prescription: policymakers should adopt a presumption that an election’s default treatment should be aligned with the result that most taxpayers would prefer. This presumption is consistent with the conclusions of tax scholars Heather Field and Emily Cauble in the American context. However, I argue that this presumption should not be ironclad: it can be overcome if structuring the election as a penalty default allows the government to glean information that can be used to improve the tax system, provided that such information cannot be obtained at lower cost by other means.

How might these default-setting principles be put into practice? To illustrate how my prescription can be applied to the status quo, I explore two provisions of the Act that contain tax elections: subsections 73(1) and 85(1). These allow a taxpayer’s transfer of appreciated property to a spouse or to a corporation, respectively, to be treated not as a realization event that triggers tax on any gains in the property but, depending on the choice of the taxpayer, as a “rollover” transaction in which tax is deferred until the transferee disposes of the property.[8] Despite their similar contexts, these two tax elections have opposite default treatments. Under subsection 73(1), the taxpayer defaults into rollover treatment. In contrast, subsection 85(1) requires that a taxpayer seeking rollover treatment affirmatively file an election.

Applying the default-setting prescription, I argue that subsections 85(1) and 73(1) should be structured consistently—with rollover treatment as the default. Evidence suggests that the majority of taxpayers who interact with both provisions prefer rollover treatment. Moreover, for those taxpayers who face high costs of compliance or who are unsophisticated about taxes, filing an election may prove too costly or burdensome to be workable. And there is little evidence that a rollover election could convey information to the government in the screening sense; taxpayers may have idiosyncratic reasons for wanting to obtain or depart from rollover treatment, and their choices in this regard would not seem to support inferences that would be useful to the tax authorities. As a result, setting the default treatment of the subsection 85(1) election as a rollover so that most taxpayers are not required to incur the costs of executing the election is preferable to the status quo alternative on both efficiency and equity dimensions.

Thus, I recommend amending subsection 85(1) in the direction of consistency with subsection 73(1) so that rollover treatment is the statutory default.[9] I am not the first to suggest this alignment: the original recommendation for subsection 85(1) made by the Carter Commission during Canada’s landmark tax reform was to confer rollover treatment on contributions of property to a corporation by default. For reasons that are not apparent from the legislative history, the default was switched during the bill-drafting process. It is not too late to rectify the error.

However, to preserve the contribution-specific disclosure that is currently provided to the government as part of the subsection 85(1) election (which does not meet my definition of “information” but which is nonetheless very important for anti-avoidance purposes), I recommend an additional policy change to accompany a switch in subsection 85(1)’s default treatment: an information reporting requirement for corporations that receive contributions of property pursuant to subsection 85(1). Rather than linking the rollover election with the government’s need for disclosure about properties contributed and consideration received in exchange, I propose to bifurcate the election from the reporting functions of the status quo by adopting an information reporting requirement that covers all contributions of property to a corporation under subsection 85(1). Both the rollover election and the government’s anti-avoidance arsenal will be stronger as a result.

This paper proceeds as follows. Part I offers a transaction cost–minimizing argument for setting tax elections’ default treatments in favour of the majority of taxpayers’ preferences. Part II examines the possibility that the reverse approach—a non-majoritarian penalty default—could yield valuable information for the government that might outweigh the costs imposed by the penalty structure. Part III documents the divergence between subsections 73(1) and 85(1) along the dimension of their default treatments. Part IV uses the default-setting principles distilled in Parts I and II to evaluate subsections 73(1) and 85(1), and presents an argument for amending subsection 85(1) to provide for a majoritarian default treatment in favour of rollover, alongside an information reporting backstop to address concerns about tax avoidance. The last part concludes.

I. The Transaction Cost Implications of Default Setting

The optimality and structure of tax elections is unexplored territory in Canada. In the American context, Heather Field is the legal scholar who has most directly addressed the general topic of elective tax provisions as a statutory mechanism.[10] In addition to Field, other scholars have examined specific US income tax elections,[11] the common law doctrine of election,[12] and issues relating to compliance with and revocation of elections,[13] but this earlier literature does not ponder the mechanism of an election as “an element of design” in the tax law, to use the title of one of Field’s articles.[14]

Most recently, Emily Cauble has suggested that if we cannot get rid of tax elections that generally disadvantage less sophisticated taxpayers, then we should structure them to increase their fairness—and she provides an excellent analysis of how to think about fairness-increasing defaults. Here, I build most directly on Cauble’s translation of default rules in the contractual setting to default rules in the tax elections setting, but reach a somewhat different conclusion. Cauble, agreeing with Field, recommends aligning an elective provision’s default treatment with the expectations of the majority of taxpayers.[15] First, a majoritarian default improves equity for those taxpayers who neglect the election or are not sophisticated enough to understand how to make the election: they are not disadvantaged by their failure to act.[16] Second, it reduces transaction costs by minimizing the number of forms that need to be filed by taxpayers and processed by the government.[17] But, on the other hand, structuring the election contrary to the expectations of most taxpayers, as a penalty default, may have a different advantage: it “may be appropriate where the government needs particular information from a taxpayer before affording the taxpayer certain favorable treatment,”[18] such as in the case of a taxpayer who must provide information about a dependent to receive the benefit of a dependency exemption. However, it remains unclear whether or how this “tagging” function of tax elections—matching eligible taxpayers with targeted tax benefits[19]—can or should defeat a presumption in favour of a majoritarian default, or why tags need to be non-majoritarian in the first place.[20] My rubric as developed below seeks to resolve these issues. Where Cauble and Field prescribe that, in general, default treatments should avoid being structured as penalties, I find that such a general presumption can be overcome where there is a measurable information benefit from a penalty default, and this measurable benefit justifies the costs imposed by the non-majoritarian structure. Further, I suggest that such measurable information benefits may be more likely to occur in settings involving corporations, where the unfairness component of the elective choice may be less pronounced.[21]

Leveraging the observation that tax elections generally increase complexity, it is clear that their existence is associated with a range of transaction costs, both for the government and for taxpayers.[22] For the government, implementing a tax election involves promulgating the necessary forms and instructions, making taxpayers aware of the election, processing taxpayers’ election decisions, providing guidance on revocability or changes in elections, and, among other administrative costs, making sure the elective treatment is applied consistently across the taxpayer’s other tax positions.[23] Like other types of transaction costs associated with taxation, these costs represent losses to society, and tax policymakers seek to minimize them.[24]

Similarly, taxpayers bear transaction costs imposed by tax elections. They can be broken down by sequence type. First, taxpayers must bear the “deliberation costs” associated with deciding whether or not to make an election. These costs may include gathering information about how the election works, understanding the default treatment, and determining whether or not the elective treatment will yield a better tax result than the default. Tax elections are likely to impose deliberation costs on all taxpayers to whom the election applies. By its nature, an election offers not just some but all taxpayers in a given situation a choice of tax treatment. The choice is offered by means of mass-distributed tax return forms and instructions, taxpayer resources, and other government communications. All of these, of course, increase the government’s costs but are designed to make taxpayers aware of the election. Certainly, there will be taxpayers who are not sensitive to available information about the explicit election because they do not absorb or pay attention to such communications. But for the taxpayers who give sufficient attention to tax information to become aware that they are confronted with a choice, they are likely to face positive deliberation costs.[25]

Second, if the result of the taxpayer’s deliberation is to proceed with making the election, there are “execution costs” associated with actually completing whatever forms are necessary to effect the election and file it with the government. The ensuing discussion in Part III of the rigours of properly filling out and filing the form for the subsection 85(1) rollover election illustrates that these execution costs can be onerous. Like the transaction costs borne by the government, deliberation and execution costs drain taxpayers’ resources without providing anything of productive value.

The transaction costs involved in deliberation about and execution of a tax election also can distort the economic decisions of taxpayers.[26] It is tempting to jump to the conclusion that a given taxpayer will incur the execution costs of making an election only if the sum of her execution costs and the deliberation costs is, in the aggregate, less than the tax savings generated by the elective tax treatment (as compared to the default treatment). However, this calculation ignores the concept of sunk costs: to determine whether to proceed with making the election, the taxpayer must first incur her deliberation costs, and those costs are sunk to gain the information necessary to know the payoff from the election. Rather than executing the election only if it is economically efficient—that is, if her benefit from the elective treatment is greater than the sum of her costs of electing—the taxpayer will make the choice of executing the election on the margin. She will elect if the execution costs are less than the benefit from the elective treatment as compared to the default treatment. Thus, the choice to execute the election on the margin may be suboptimal from an efficiency perspective, and can be seen as a welfare-reducing distortion of taxpayer behaviour.

In light of these transaction costs imposed by tax elections, it is clear that the default treatment of a tax election matters. While there may be some scope for the government to decrease taxpayers’ deliberation costs, such as by providing information that will help taxpayers decide quickly whether they should or should not make the election, minimizing execution costs is an obvious way to make the election less costly. Assuming for the moment that transaction cost minimization is the government’s sole objective, then the default treatment for the election should be set in a majoritarian fashion—that is, so that the majority of taxpayers will not desire to elect out of the default. As a result, only a minority of taxpayers will need to incur execution costs, and transaction costs will be reduced.

In addition to the straightforward transaction cost story, there is another reason why default-setting matters: well-documented behavioural patterns that result in defaults being “sticky”. The literature demonstrates that agents are often reluctant to opt out of a default despite it being advantageous for them to do so; people tend to stick with their current situation, even when faced with a choice.[27] This has been dubbed the “status quo bias”, and can amplify the stakes of setting the right default rules.[28]

II. Setting Defaults to Reveal Information

This Part considers the possibility that minimizing transaction costs may not be the sole dimension along which to evaluate how to set the default treatment of a tax election. There is considerable literature on default rules in the contexts of contract law, commercial transactions, and corporate law. Much of it arises in relation to how courts should fill the gaps where contracts are left incomplete by the parties but need to be interpreted to resolve disputes. Some scholars advocate that courts should exercise this gap-filling function by interpreting the contract in a majoritarian fashion—that is, filling gaps with the terms that the majority of parties would have wanted had they considered the issue ex ante. Where the source of contractual incompleteness is the transaction cost of negotiating the full range of possible contingencies in a contract, then majoritarianism has obvious appeal—the default rule can be set such that most parties do not have to expend the costs of contracting around it.

But other commentators have criticized this approach, arguing that majoritarianism oversimplifies the calculation. Ian Ayres and Robert Gertner show that choices in the law can operate to induce individuals to self-select along different dimensions or characteristics.[29] In particular, there are

[t]wo related mechanisms for inducing actors to reveal private information [that] are well understood. One solution is to impose a punitive default and allow individuals to elect out of it in a way that produces information about them. Another approach is to force actors to choose from a menu of options structured in such a way that an actor’s choice discloses salient information about his preferences. In either case, agents are compelled to choose a course of action and their choices produce information sought by the principal. Agents’ choices are revealing.[30]

However, the idea that punitive defaults can induce actors to self-select has only fairly recently gained traction in the tax context.[31] Recent work by Benjamin Alarie argues that the government routinely acts as a second-degree price discriminator[32] by harnessing self-selection among taxpayers in the context of anti-avoidance enforcement efforts.[33] And Alex Raskolnikov has proposed a “dual enforcement system” in which taxpayers are forced to choose between two regimes, with the objective of allowing the government to better target tax enforcement to taxpayers’ diverse motivations for paying or not paying taxes.[34]

In other work, I argue that elections can act as screens to separate taxpayers according to their costs of complying with the election.[35] Here, Raskolnikov’s dual enforcement system concept is relevant because it illustrates how explicit elective provisions in the tax law can cause taxpayers to separate themselves along characteristics that are unobservable but valuable for effective tax administration. Indeed, most of the inefficiencies inherent in the tax system would disappear if the government possessed perfect information about taxpayers’ characteristics and preferences: first-best lump sum taxation would be possible, enforcement techniques could be finely tailored, outreach could be targeted to increase salience, and other tax miracles could be performed. Raskolnikov proposes maintaining our current “deterrence regime” with high statutory fines and well-publicized punishments alongside a newly developed “compliance regime” with lower statutory fines, an emphasis on taxpayer service, and standards of review that increase the probability of conviction.[36] Under his proposal, taxpayers would choose one of two options from a menu of regimes, such that “gamers” predominantly would find it advantageous to choose deterrence and “non-gamers” compliance. To the extent that taxpayers’ choices allow the government to separate gamers from the rest of the taxpayer population, tax enforcement efforts could be more efficiently targeted to take into account taxpayers’ diverse motivations, achieving higher compliance at lower social cost. Raskolnikov’s proposal relies on a menu of choices being presented to taxpayers, but its principle can be extended to the use of a non-majoritarian default that acts as a “penalty” for most taxpayers. The penalty would force them to reveal their enforcement preferences by electing out of the default regime.[37]

These two sets of rationales for setting elective tax provisions’ default treatments point in opposite directions—a majoritarian default can minimize transaction costs but a non-majoritarian default can reveal valuable taxpayer information to the government. Each rationale can reduce the costs involved in having an income tax. In order to determine which consideration dominates, policymakers will need to ask questions such as: How onerous are the transaction costs imposed on taxpayers by the need to execute the election? And what, if any, role would be played by the information revealed by a non-majoritarian penalty default?

While it is hard to avoid the fuzzy conclusion that the facts and circumstances of the particular election will determine policymakers’ balancing between the two rationales, it bears noting that the two intersect in an important way. Where the government seeks to gain information through its choice of a penalty default, the execution costs associated with the election may blunt any information-forcing function that the election might possess, in the following way: Taxpayers make the choice to elect out of the default treatment on the margin. Therefore, if the tax benefits from electing out are swamped by the execution costs of the election (e.g., a complicated form or onerous rules for filing), taxpayers will choose rationally to simply stick with the default. The “sticky default” behavioural bias will exacerbate this problem. In the case of a penalty default with high execution costs, only a small subset of taxpayers who would otherwise choose to elect out of the penalty will find it worthwhile to do so. As a result, where execution costs are high, the presumption in favour of a majoritarian default becomes stronger as the burden on the government of showing how the penalty default will reveal information about the specific electing-out cohort of taxpayers becomes heavier.

III. The Opposite Defaults of the Subsection 73(1) and 85(1) Elections

In exploring why there is a divergence in the default treatments prescribed under the subsection 73(1) and 85(1) elections, it makes sense to first take a brief tour of the mechanics and legislative histories of each subsection. The following Part shows that, notwithstanding the technical differences in the applications of rollover treatment between the two subsections, their underlying policy motivations are the same and thus are unlikely to explain the divergence in default treatments.

A. Why Offer a Rollover in the First Place?

Subsections 85(1) and 73(1) are associated with the most significant tax reform in Canada’s history.[38] The Royal Commission on Taxation, or the Carter Commission (after its chair, Kenneth Carter), was appointed in 1962 and cast particularly intense scrutiny upon the base on which the income tax was assessed. Under the 1952 Act and its predecessors, capital gains—increases in the value of capital property as distinguished from the income streams generated by such property—were not considered part of taxable income at all.[39] The Carter Commission recommended including capital gains in the income tax base due to considerations of fairness, efficiency, and progressivity.[40] The subsequent adoption of legislation embodying this policy shift is the Commission’s most important legacy.

Despite the compelling justifications for broadening the income tax base to include capital gains, the new policy raised thorny questions about the timing of taxable events. The Carter Commission accepted the need for a tax system that was rooted in the “realization principle”—the rule that tax should only be imposed upon a specified triggering event. Typically, a realization event would generate liquidity with which to pay the resulting tax liability, such as a sale of a capital asset for cash. However, to minimize the opportunities for taxpayers to avoid realization events and as a second-best approximation of an ideal accrual-based income tax, the Commission recommended that “the term ‘disposition’ should be used in the broadest sense.”[41] Indeed, the statute as adopted and as it stands today reflects this breadth, covering almost every transfer so long as the transaction results in a change of beneficial ownership of the property (but even this is not required for some dispositions relating to trusts).[42] Events that are not transfers may nevertheless constitute dispositions for purposes of the Act; neither does the taxpayer need to actually receive the proceeds of the disposition for a disposition to have occurred. For instance, when a debt or share is redeemed, cancelled, or converted, or an option expires, a disposition has occurred and tax—whether or not an individual has the cash to pay it—may be due.[43]

As a result of the new tax policy relating to capital gains, taxpayers faced incentives to avoid tax-triggering transfers of capital property. Commentators have called this the “lock-in effect”.[44] Particularly in cases where the person or entity holding legal title to the property changes but the beneficial ownership remains within a single economic unit (such as corporate stakeholders or members of a household), policymakers were concerned that taxing capital gains on disposition might distort taxpayers’ asset allocation behaviour and, in turn, damage Canada’s economic competitiveness. Tax-deferred rollover treatment for property transferred to a spouse or contributed to a corporation can be seen as a departure from the strict application of the realization principle designed to address concerns about the lock-in effect, as shown in the subsection-specific discussions that follow.[45]

B.The Subsection 73(1) Rollover

1. Mechanics

If subsection 73(1) did not exist, a transfer of appreciated property from one spouse to another generally would result in a tax liability for the transferor, and the transferee spouse would take the property with a cost base equal to fair market value.[46] Subsection 69(1) articulates the general rule for non-arm’s-length transfers of property, such as those between spouses: they are treated as taking place at fair market value, regardless of the actual consideration given in exchange for the property.[47] This deeming rule is designed “to prevent manipulation of asset transfer prices between parties who are not dealing with each other on an economic arm’s length basis.”[48] Subsection 69(1) further prescribes that any resulting capital gain, or any allowable capital loss not limited by the stop-loss rules,[49] is recognized by the transferor.[50]

To avoid this result, subsection 73(1) provides a deeming rule that carves out qualified transfers of property between spouses from the reach of subsection 69(1). It deems a qualified transfer as occurring, generally, at “proceeds equal to ... the adjusted cost base to the [transferor] individual of the particular property immediately before [the time of the transfer].”[51] By treating the proceeds of disposition as equal to the transferor’s adjusted cost base, the transferor realizes no taxable capital gains as a result of the transfer.[52] The transferee inherits the adjusted cost base from the transferor, and tax is deferred until the property is disposed by the transferee in another transaction.[53] In addition, sections 74.1 and 74.2 generally operate in concert with subsection 73(1) to attribute any income or losses from the transferred property (or substitute property) back to the transferor, rather than attributing them to the transferee spouse. These attribution rules also apply to any taxable capital gains or allowable capital losses generated by the transferred property.[54]

Rollover treatment for qualifying spousal transfers of property, therefore, is the statutory default: subsection 73(1) mandates rollover treatment “unless the individual elects in the individual’s return of income under this Part for the taxation year in which the property was transferred that the provisions of this subsection not apply.”[55] In the event that the transferor spouse makes the election, the transfer is treated as a disposition of property at fair market value pursuant to the deeming rule of subsection 69(1). Furthermore, subsection 74.5(1) provides that, where the transferor files an election under subsection 73(1) and the fair market value of the transferred property did not exceed the consideration given in exchange for it at the time of transfer, the attribution rules of sections 74.1 and 74.2 are suspended.[56] Thus, rollover treatment accompanied by the spousal attribution rules is the clear default, and electing out of the default treatment deems the transfer to have taken place at fair market value consideration for the transferred property, even if the consideration was less than fair market value. If the taxpayer does not elect out of the default, the income, gain, and loss attribution rules will continue to apply.[57]

According to Interpretation Bulletin IT-325R2, the mechanics of making the election are trivial: “There is no official form for the election not to have the subsection 73(1) rollover apply. This election is normally made by the transferor simply reporting the full tax consequences of the disposition on his or her Income Tax Return for the year of the transfer.”[58] However, additional rules, such as whether the election can be revoked via an adjustment to the transferor’s income tax return, are not readily available.

2. Statutory History

As Justice Bowman stated in Lipson v. Canada, “[s]ubsection 73(1) has as its purpose the facilitation of inter-spousal transfers of property without immediate tax consequences.”[59] And Neil Brooks has characterized the purpose of the rollover provision as reflecting the economic reality of shared property within a family: “Within marital units it is often difficult to determine the beneficial ownership of property. ... Providing a rollover for interspousal transfers eliminates what would be a serious tracing problem.”[60] Both interpretations are consistent with the larger rationale of avoiding capital property lock-in, particularly because the individual—and not the family—remains the unit of taxation in Canada, despite the Carter Commission’s recommendation to the contrary.[61]

The subsection 73(1) rollover provision was included as part of the 1972 tax reform bill, but it did not contain an election. Instead, rollover treatment for spousal transfers of property was mandatory. Only beginning in 1979 was the transferor spouse offered an election to avoid the application of subsection 73(1)’s rollover treatment.

It is not clear from the legislative or judicial history why the election was added, although one may speculate that it may have been to avoid complications where property was being transferred in the context of marital breakdown. For instance, when spouses are in the process of separating, a transfer of property absent the election would saddle the transferee spouse with future taxes on any built-in gain associated with the property, as well as the income from the property once the section 74.1 and 74.2 attribution rules ceased to apply. Unless the couple was savvy enough to negotiate an intermediate transfer at arm’s length to a third party to obtain a cost base of fair market value, default rollover treatment could cause complications in splitting up property or, at worst, cause an unwitting transferee spouse to face an unexpected tax liability after the dissolution of the marriage became final. Alternatively, the adoption of the election may have had, at the time, an expressive equity rationale. Allowing spouses to choose to treat a transfer within the couple as a “real” transaction that occurred at arm’s length may have conveyed respect for the idea that spouses could be treated as economic equals, reflecting changing attitudes toward gender and family.

C. The Subsection 85(1) Rollover

1. Mechanics

Like subsection 73(1), subsection 85(1) operates as an exception to the subsection 69(1) deeming rules and applies to contributions of property to a corporation under certain conditions.[62] If these conditions are met, the taxpayer and the transferee corporation may file a joint election[63] specifying an “elected amount” that is deemed to constitute, generally, three things: the taxpayer’s proceeds of disposition,[64] the taxpayer’s cost of the consideration received from the corporation,[65] and the corporation’s cost in the transferred property.[66]

Paragraphs 85(1)(b), (c), (c.1), (d), and (e.1) contain the basic deeming rules[67] that operate to constrain the value that the taxpayer and the corporation may jointly elect within a certain range—that is, they designate upper and lower bounds on the elected amount.[68] This structural feature distinguishes it from subsection 73(1): for spousal transfers, the election provides a binary option—either rollover treatment or disposition at fair market value. By contrast, subsection 85(1) can be seen as offering a menu of elective treatments because taxpayers can jointly elect any amount between the applicable upper and lower bounds.

The admittedly complicated upper and lower bound rules on the elected amount have two key implications: First, a taxpayer who transfers property that has an accrued loss to a corporation must realize that loss upon transfer.[69] The discernible policy rationale behind this outcome is to prevent corporations from importing losses from their contributing shareholders; this outcome represents an important difference as compared to straight nonrecognition treatment. Second, in the event that the upper and lower bound provisions conflict and the taxpayer receives a subsection 15(1) benefit that is included in taxable income, the amount of the benefit will increase the taxpayer’s cost in the consideration received from the corporation pursuant to subsection 52(1).[70]

Unlike filing an election under subsection 73(1), which can be done on the transferor’s annual tax return, the rules for making a valid election or a valid series of elections under section 85(1) are far more complicated. Form T2057 is required to make the subsection 85(1) election;[71] on this form a box dubiously labeled “informative notes” reminds taxpayers that “[t]he rules for section 85 elections are complex. ... If this form is incomplete, the Canada Revenue Agency may consider the election invalid, and subsequent submissions may be subject to a late-filing penalty.”[72] Commentators emphasize that the election is not for the faint of heart,[73] and that “the slightest deviation from the technical rules of the section can invalidate the rollover and trigger a deemed disposition at fair market value.”[74]

Two particular aspects of the subsection 85(1) election increase its complexity. First, the election must be made jointly by the transferor taxpayer and the transferee corporation.[75] Under subsection 85(6), the election must be “made on or before the day that is the earliest of the days on or before which any taxpayer making the election is required to file a return of income pursuant to section 150 for the taxation year in which the transaction to which the election relates occurred.”[76] Given the potential difficulties involved in making sure that the parties communicate about their tax return due dates and finalize the rollover form in advance of its due date, it is not surprising that there are remedial provisions for late filings: a subsection 85(1) election can be filed up to three years after the original deadline under subsection 85(6).[77] However, there is a penalty.[78]

Second, the election must be made on an asset-by-asset basis.[79] Because the cost basis of the contributed property must be properly allocated among the shares and any boot received in exchange, and the tax consequences can differ in material respects depending on the order in which different properties are contributed,[80] the election requires a “specific and adequate description of each property transferred,” including the fair market value, adjusted cost basis, agreed amount, and information about the consideration received in exchange for the property.[81] These details are required for each property, with variations depending on the type of property contributed.[82] The taxpayer must retain all supporting schedules and records necessary to substantiate her position.

2. Statutory History

Why does subsection 85(1) exist? The Carter Commission offered a spirited defense of allowing rollover treatment for contributions of property to a corporation as a way of combating the lock-in effect:

We recognize that it is often necessary to change the form of ownership of a business or property, or to rearrange or reorganize the affairs of corporations for business reasons. If every such change or reorganization were to result in a disposition for tax purposes by the shareholders, or the corporation, or both, this could have an inhibiting effect and could tend to produce undesirable rigidity in corporate structures. Because we regard a corporation as an intermediary, and individuals as the persons who ultimately bear the taxes, we consider that certain corporate reorganizations and transfers which change the form of ownership, but do not effect a change in the ultimate beneficial ownership of a business or property, should not result in a tax liability.[83]

The Commission was surely aware that subchapter C of the Internal Revenue Code of the United States confers nonrecognition treatment on contributions of property to a corporation where the contributor owns a requisite majority of the corporation’s shares after the contribution.[84] Although the Commission did not indicate whether its recommendations were inspired by its neighbour’s approach, its resulting prescription took a similar tack[85] and proposed that contributions of property to a corporation should, in general, not trigger tax.[86]

Furthermore, the Commission added a twist not present in the corresponding American provision: it suggested that taxpayers who wanted partial gains to be recognized upon contributing property to a corporation, and their cost bases in the property and the shares adjusted accordingly, should be able to choose recognition treatment. In particular, it proposed that the law allow taxpayers to elect out of the default of rollover tax treatment.[87] In recommending that a taxpayer be allowed to opt out of rollover treatment, the Commission can be seen as affirmatively encouraging taxpayers to make contributions of property to corporations—regardless of their particular tax situations, which could make rollover or realization treatment more or less advantageous.

Consistent with the recommendations of the Carter Commission, the bill that was introduced in the House of Commons duly included an elective rollover provision for contributions of property to a corporation.[88] However, a curious thing happened between the time that the Carter Commission report was released and commented upon by the Minister of Finance in a White Paper: the direction of the election for rollover treatment upon contribution of property to a corporation was reversed. Rather than drafting the legislation such that rollover treatment was the default, the legislation allowed the opportunity for tax deferral only “if the taxpayer and the corporation have jointly so elected in prescribed form and within prescribed time.”[89] As summarized by the practitioners who published the CCH Explanation of the legislation, “a joint election must be filed in order to take advantage of the roll-over provisions. If no election were filed, the rules in subsection 69(1) regarding inadequate consideration would presumably apply.”[90]

No available sources indicate why the default tax treatment was reversed in the legislation.[91] Although numerous speeches were made by Members of Parliament and others,[92] a review of the floor debates in the House of Commons yielded no mention of the use of elective provisions in general, of the rollover treatment of property upon contribution to a corporation in particular, or of why the election default was reversed.[93]

Which default rule—if either—is superior? And is it the same for subsection 73(1) as for subsection 85(1)? The following Part addresses these questions in light of the prescription developed in Parts I and II.

IV. Evaluating Subsections 73(1) and 85(1)

Here, we can see the prescriptive rubric for evaluating tax elections’ default provisions in action: an election’s default treatment generally should align with the result that most taxpayers would prefer. However, this presumption can be overcome if structuring the election with a non-majoritarian penalty default would allow the government to glean valuable and otherwise unobservable information about taxpayers that could be used to increase the efficiency of the tax system.

For each of the rollover provisions, I first evaluate whether the status quo election is structured as a majoritarian default and how this structure influences the execution costs at stake. Second, I ask whether it is plausible that structuring the default treatment as a penalty in order to cause some taxpayers to elect out might reveal valuable information to the government.

A. Subsection 73(1)

The default structure of subsection 73(1) appears to be majoritarian. Perhaps because it did not start off as a rollover provision that included an explicit election, its current structure seems to accommodate most taxpayers’ preferences to defer tax on any accrued gains in capital property transferred to a spouse. While it would be helpful to have survey evidence or other data about taxpayer preferences for rollover treatment in the context of spousal property transfers, it seems fair to assume that most taxpayers would prefer to defer any gains on appreciation in the transferred property, even if this causes the future income from the property to be attributed to them. This conclusion results from the following risk-benefit analysis: the taxable capital gains on the property at the time of transfer are certain, while the future income stream is speculative.

Of course, particular taxpayers may have idiosyncratic circumstances that would make electing out of rollover treatment and adjusting the cost basis to fair market value attractive. First, they may have allowable capital losses that could offset realized gains. Second, if the transferor spouse was in a significantly higher bracket than the transferee spouse and the capital asset was expected to generate significant income streams in the future, the impact of circumventing the section 74 attribution rules might be worth the tax liability on an immediate realization of gains. Third, in a similar disparate-bracket situation, electing out of rollover treatment can facilitate income splitting in the following way. When (as now) the prescribed rate for spousal loans is low, the higher-earning spouse can make a loan to the lower-earning spouse at the prescribed rate of interest.[94] At the same time, the higher-earning spouse can transfer income-earning property (assets such as stocks, bonds, etc.) to the lower-income spouse at fair market value and circumvent attribution by making the proper election. The lower-income spouse can deduct the interest expense since the loan was incurred to earn income.[95] When the dust settles, the higher-income spouse has interest income (plus the capital gain or loss, if any, on the property that was transferred), but this income is likely to be more than offset by foregoing the income on the property that was transferred to the lower-earning spouse. Finally, as discussed above, where spouses are separating before the final dissolution of their marriage, electing out of the default treatment may simplify the division of property by waiving the attribution rules and accelerating tax on any gains.[96]

Notwithstanding this majoritarian default structure, it bears noting that the subsection 73(1) election has a fairly low impact on execution costs. The fact that there is simply a box to check on the individual tax return, without further reporting requirements, implies that executing the election is not difficult. While deliberation costs may not be negligible—all taxpayers who transfer property to a spouse are confronted with the choice about whether to treat the transfer as rollover or not—the execution costs that would be minimized by a majoritarian election appear to be already low.

The low execution costs of the 73(1) election reduce the stakes of default-setting. Where execution costs are low, imposing them on a majority of taxpayers—via a penalty default structure—would have a smaller impact than if execution costs were higher. If forcing taxpayers to elect out of fair market value realization treatment and into rollover treatment would reveal valuable information to the government, one would expect that the costs of checking a box would not deter taxpayers from making the election and conveying the valuable information.

Does a plausible story exist of how a penalty default structure in which taxpayers were required to elect into rollover treatment would reveal information to the government? Anything in this regard is speculative, and penalizing taxpayer inattentiveness certainly seems to be a strategy for generating revenue at low political cost.[97] But it is harder to identify an account of how structuring the subsection 73(1) election as a penalty default might reveal taxpayer information that the government could leverage to increase the efficiency of the tax system. Moreover, instituting a penalty default in this case would run the risk of fostering the undesirable outcome of capital lock-in. To the extent that the penalty default proved sticky or, for whatever reason, a taxpayer who would benefit from rollover treatment failed to elect in, the clear policy objective of making rollover treatment available in the first place would be defeated. Therefore, applying the default-setting rubric developed above yields a clear conclusion: subsection 73(1) is properly structured as a majoritarian default, and there is insufficient justification for reimagining it in a penalty-default structure.

B. Subsection 85(1)

Applying the default-setting rubric to subsection 85(1) yields a more rousing conclusion. Starting with the 1972 Act, the election’s default treatment in favour of fair market value realization appears to have been set not in accordance with the preferences of the majority of taxpayers but rather as a penalty. True, the election as it is structured delivers a variety of useful disclosures to the government—about the cost basis of contributed property, its effect on paid-up capital, and other dimensions that allow the government to police tax avoidance. But this information can be elicited in a lower-cost and more effective manner than by structuring the election as a penalty for the majority of taxpayers. As a result, I suggest that Parliament should amend subsection 85(1) to adopt rollover treatment as the default, consistent with the original recommendations of the Carter Commission, and should pair this amendment with a new mandatory reporting requirement for corporations receiving property pursuant to subsection 85(1).[98] I discuss each policy reform in turn.

That the current default structure of the subsection 85(1) election is non-majoritarian should not provoke controversy. Practitioners confirm that most taxpayers contributing property to a corporation under 85(1) elect straight rollover treatment rather than an intermediate elected amount or the realization default (transfer at fair market value). Such anecdotes should be verified by systematically surveying tax filings or taxpayers themselves, but assuming that the distribution of taxpayer preferences in favor of rollover is sufficiently lopsided, the status quo can be seen a penalty default that imposes execution costs on a majority of taxpayers.

Moreover, the hefty execution costs of electing out of subsection 85(1)’s default treatment raise the stakes of getting the default right. Making an election requires coordination between the transferee and the corporation; a special form must be filed before the prescribed deadline; each contributed property requires a separate election; and so on. To the extent that a majority of taxpayers elect out of the default treatment, there is scope for reducing transaction costs. Still, under the default-setting rubric, the presumption in favor of a majoritarian default can be overcome if structuring the default as a penalty yields sufficiently valuable information to justify the penalty’s associated increase in costs. Is this the case for subsection 85(1)? I contend that it is not, because the reasons that taxpayers may decide not to seek rollover treatment in the context of subsection 85(1) are so varied. For instance, rollover treatment may be unfavourable where the transferor has accrued capital losses carried over from previous periods. Because allowable capital losses can only absorb taxable capital gains, it makes sense to use them as soon as possible.[99] Second, for taxpayers disposing of qualified small business corporation shares by contributing them to another corporation pursuant to subsection 85(1), the capital gains exemption confers tax-free treatment on a threshold amount of gains without needing to make an election.[100] Third, parties may want to avoid having two properties—both the asset held by the corporation and any shares received in exchange—with low cost bases.[101] The rollover staves off tax that would be owed by the transferor upon contribution of the property, but it is a deferral of, not an exemption from, future tax liability. Because the rollover deprives both the transferor (in the consideration received for the property) and the transferee (in the property transferred) of a step up in cost basis to fair market value, a rollover generates two properties with low cost bases in place of just one. When either party disposes of her property, tax will be due—both the transferor will pay tax on the gains in the shares and the corporation will pay tax on the gains in the property. Of course, the severity of this disincentive will vary with the expected amount of time that the transferor and the transferee intend to hold their respective properties. If the expected holding period is long, deferral may be more attractive.[102] The literature mentions various other situations in which rollover treatment would not be preferable, but all seem unlikely to apply to the circumstances facing the vast majority of taxpayers.[103]

Because there are so many idiosyncratic reasons that taxpayers might decide to elect out of rollover treatment (even if it is a minority of taxpayers that choose to do so), the potential for the election to act as a useful screen is dampened. The government cannot make clear inferences on the basis of a particular taxpayer’s or corporation’s decision to elect or not elect rollover versus fair market value versus intermediate elected amounts in the context of subsection 85(1). As a result, it is very unlikely that the election can produce information sufficient to rebut the presumption in favour of a majoritarian default. In any event, an information-forcing function of a non-majoritarian election would have to be identified by the government and debated internally, during which (presumably) the benefits of the information yielded by the election would be weighed against the transaction costs imposed by a penalty default. Given the lack of clues or a sufficiently compelling story, I conclude that the presumption in favour of a majoritarian default should stand.

There is also an equity dimension that must be considered alongside the efficiency rubric that I offer. Because the transaction costs of electing out of the default treatment—however it is set—are substantial, some taxpayers will be deterred from making the election. The deterred taxpayers are likely, on balance, to be less sophisticated or have access to fewer resources relative to taxpayers making the election. While this may be less true for subsection 85(1) as compared to subsection 73(1) because contributions of property to a corporation are likely to be made by more sophisticated taxpayers or those represented by counsel or accountants, the execution costs will undoubtedly affect taxpayers on the margin. Where the default treatment is majoritarian in that it meets the preferences of most taxpayers (including, in particular, those at the bottom of the resource spectrum), the effect of the provision will be more equitable than if it was structured with a penalty default.

Merely switching the subsection 85(1) default to rollover treatment is not a silver bullet. In order to preserve the current statute’s flexibility as well as its policy of preventing importation of losses, rollover treatment would apply by default only when the taxpayer’s cost exceeded the fair market value of the property. Otherwise, the transferee corporation could receive property that, if disposed after transfer, would yield a capital loss. And, to continue to offer taxpayers the full range of elected amounts, the elective treatment under an amended statute would need to allow for the same range as under current law. Finally, where property was contributed in exchange for both share and non-share consideration, taxpayers would still need to allocate the cost among the shares and any boot received in exchange. Thus, switching the default treatment to rollover is not a panacea for the complexity inherent in subsection 85(1). But avoiding a situation where a majority of taxpayers find themselves scrambling to file an extremely complicated election before a deadline in order to avoid penalties would be a substantial step in the direction of increased efficiency and equity.

Separate from the case for a majoritarian default treatment for subsection 85(1), however, is the concern that changing the status quo default treatment would have unintended consequences in other areas of tax administration, particularly anti-avoidance enforcement. The current penalty default structure of the subsection 85(1) election also requires substantial disclosure from taxpayers. Form T2057 requires electing taxpayers to provide many details about the property contributed (its fair market value, adjusted cost base, and other details) and the share consideration received in exchange for the contributed property (a description of shares, their paid-up capital, and the fair market value of total consideration received).[104] This disclosure is used by the government to make a number of determinations that are central to its efforts to combat tax avoidance,[105] including: whether the paid-up capital of the share consideration issued in exchange for the property might result in a deemed dividend,[106] whether the indirect gift rule is applicable,[107] whether a paid-up capital “grind” is required for the shares issued as consideration,[108] and other determinations central to properly assessing taxes due. Because this disclosure about contributions of property covered by subsection 85(1) is so important in allowing the government to combat surplus-stripping and other tax avoidance strategies, I propose the adoption of a mandatory reporting requirement for all disclosure currently provided on Form T2057.[109] Best of all, such reporting would not depend on an election being filed, and thus would not deprive the government of disclosure simply on the basis of an idiosyncratic elective choice. To the extent that the anti-avoidance gains justify the costs that such a requirement would impose on reporting corporations and their shareholders, it should be adopted as a complement to switching the default treatment of the subsection 85(1) election.

Conclusion

This paper accomplishes two things. First, it builds a principled rubric to assist policymakers in thinking about how to set the default treatment for tax elections. The rubric balances the transaction costs of executing tax elections against the possibility that taxpayers’ elective choices can yield valuable private information to the government. I show that, in general, a tax election’s default treatment should align with the result that most taxpayers would prefer. However, this presumption can be overcome if structuring the election as a non-majoritarian penalty default would allow the government to glean valuable information about taxpayers that could be used to improve the tax system, provided that such information is not available elsewhere at lower cost.

Second, this paper analyzes an inconsistency in the structure of two key tax elections in the Act relating to contributions of property to a spouse and to a corporation, respectively. Whereas subsection 73(1)’s default treatment is rollover, subsection 85(1)’s default treatment is fair market value realization, despite the recommendation of the Carter Commission to the contrary. When the default-setting rubric is applied to the subsection 73(1) and 85(1) rollover provisions, the results are clear. Subsection 73(1)’s default is likely aligned with majoritarian preferences, and in any event the execution costs of making the election are low and there does not appear to be an information-forcing purpose in adopting a non-majoritarian penalty default. Thus, I recommend no change in the structure of subsection 73(1). On the other hand, subsection 85(1) was structured, without explanation, contrary to the recommendations of the Carter Commission, thus saddling taxpayers with a default that most will be forced to elect out of. The costs of executing the election are not trivial. Moreover, there is not a persuasive story according to which a penalty default would cause certain taxpayers to reveal information about themselves to the government. Therefore, I recommend a simple change to the status quo subsection 85(1) election: rollover treatment should be the default treatment, consistent with the subsection 73(1) structure as well as the initial recommendations of the Carter Commission.

Such a change to the default of subsection 85(1) would not be earth-shattering in its impact on the tax system, but it would undoubtedly decrease pressure on taxpayers to elect out of a penalty default and thereby reduce aggregate execution costs. Moreover, it would have desirable equity effects—in the event that a taxpayer was not represented by counsel or was unsophisticated about the tax consequences of making the election, her inaction would have less severe consequences. If such a default-switch were accompanied by a mandatory reporting requirement for corporations receiving contributions of property pursuant to subsection 85(1), the possible negative consequences to anti-avoidance enforcement of changing the subsection 85(1) election could be mitigated. Reimagining the default structure of subsection 85(1) would be a small but meaningful step in the right direction for Canada’s income tax system.