Introduction
A fundamental principle of Westminster democracy is parliamentary control over public finance: “Responsibility for the public purse in parliamentary democracies lies with parliament, which enjoys constitutional or conventional control over revenue and expenditure” (Wanna, Reference Wanna and Weller1997: 148). How meaningful this control is has long been debated. A.V. Dicey saw it as a crowning achievement of parliamentary sovereignty, while Walter Bagehot saw it as “more constitutional fiction than fact” (Bateman, Reference Bateman2020: 10). It is a principle of Canadian government that parliamentary authorization is required for expenditure by the executive (Ward, Reference Ward1964: 3). Authorization to spend is typically given in advance (budgetary estimates) and for a specified purpose (votes). The object of study in this article—special warrants—is a spending instrument where authorization is granted by parliament retroactively and spending is for an unforeseen purpose. Special warrants trade off a reduction in parliamentary control for an increase in executive velocity.
It is claimed that liberal democracies face challenges adapting to a higher-speed world while maintaining the separation of powers. Whereas legislatures need time for debate and courts need reflection for judgement, the executive is compelled to act in ever-tightening time frames (Scheuerman, Reference Scheuerman2004). This is no mere esoteric distinction: Some believe “modern bureaucratic conditions make strict compliance with public finance law, particularly appropriations legislation, a practical impossibility” (Bateman, Reference Bateman2020: 211). The rise of the “Schmittian administrative state” leads to legal “grey holes,” where there are “some legal constraints on executive action—it is not a lawless void—but the constraints are so insubstantial that they pretty well permit government to do as it pleases” (Vermeule, Reference Vermeule2009: 1069; on “grey holes” see Dyzenhaus Reference Dyzenhaus2006). It has thus become a “significant responsibility of the executive to restrain itself” (Health, Reference Health2020: 298).
Scholars in Canada and elsewhere are dubious of executive self-restraint, frequently bemoaning “increasing centralization in the hands of the executive” (Thomas and Lewis, Reference Thomas and Lewis2019: 363). When not constrained by law, executives can be restrained by convention. Constitutional conventions help to control the manner in which power is exercised. Research on constitutional conventions is largely about their nature (Dicey, Reference Dicey1915; Jennings, Reference Jennings1959), their function (Marshall, Reference Marshall1984; Heard, Reference Heard1991; Forsey, Reference Forsey2020) and the way in which they bind (Jaconelli, Reference Jaconelli1999; Sirota, Reference Sirota2011; Albert, Reference Albert2015). Rarely is research dedicated to the history of a specific convention (Brock and Bowden, Reference Brock and Bowden2024; Miragliotta and Algra-Maschio, Reference Miragliotta, Algra-Maschio and Barry2025). Historical analysis matters because constitutional conventions have to be more than notional: “Customs and conventions arise from what people do, not from what they agree or promise” (Jaconelli, Reference Jaconelli2005: 169). As such, we need to understand governing in practice to know whether conventions are indeed commanding restraint in accordance with constitutional principles.
This research note employs special warrants as a case study to examine how constitutional controls over executive power differ in law, convention and practice. Special warrants are a mechanism to enable executive spending absent typical parliamentary approval. Two examples are: (a) the legislature is not sitting and cannot readily be recalled, but there is urgent need to act (for example, disaster relief), or (b) a general election coincides with the end of the fiscal year and the expiry of the previous year’s expenditure authorization, but funding is needed to provide regular public services until a legislature is returned and a budget passed. Special warrants are issued under statute, with the legal conditions for their use differing among Canada’s jurisdictions. Special warrants are to be used only when parliamentary authorization cannot be obtained but exigencies of state urgently require expenditure. Evidence in this research note indicates that special warrants are increasingly being used as a convenience. Does the modern use of special warrants constitute a change in Canadian constitutional convention?
We advance three main points. The first is that the laws, conventions and practices that shape the separation of powers in Canadian government result in real trade-offs between control and speed. Historically, it was conceived that “preventing unilateral executive action will increase state credibility and revenues but reduce the state’s capacity to react quickly to events” (Cox, Reference Cox2016: 178). The speeding up of public administration, not just in response to crises such as war or pandemic but in response to citizen expectations in a continuous state of exception, make special warrants a mechanism potentially attractive to executives to actually increase fiscal capacity and government credibility.
Second, in political science it is important to apply historical perspective to analyze Canadian governmental institutions and processes (Lucas and Vipond, Reference Lucas and Vipond2017). We agree with Bateman that, “historical analysis reveals the pathway upon which the modern design of financial constitutionalism depends” (Bateman, Reference Bateman2020: 19). Historical analysis reveals that the contemporary use of special warrants is less likely to adhere to convention, but also that this phenomenon is more long-standing than is commonly understood. Using historical analysis and a compendium of legislative provisions, we use special warrants as a case study in how our understanding of public administration, based on law and convention, can be reshaped by practice.
Third, attention to special warrants is timely, as this has not been done in over 60 years,Footnote 1 and yet, it is again attracting attention from legislatures and media. We seek to provide an up-to-date understanding of this aspect of Canadian parliamentary constitutionalism and public finance that has received scant attention,Footnote 2 particularly in light of increased use of special warrants for routine spending. Throughout, we consider law, convention and practice, structuring this research note to examine special warrants through each lens, recalling the classic work of Graham Allison to “demonstrate how alternative conceptual lenses lead one to see, emphasize, and worry about quite different aspects” (Allison, Reference Allison1971: v).
Law of Special Warrants
Parliamentary control over executive expenditure was affirmed in the Glorious Revolution of 1688 (Maitland, Reference Maitland1908: 247). In the decades that followed, parliamentary authority was formalized through annual budgets; executive authority moved from the monarch to a cabinet responsible to Parliament. “The first step,” explains Gary Cox, “was to entrench [rules regarding expenditure] in statute. The next step was to remove all executive discretion in interpreting statutory provisions” (Cox, Reference Cox2016: 26). As the process of annual budgets matured, parliamentary control over expenditures adapted, and by means of the “confidence convention,” Parliament and the executive spoke together (Ward, Reference Ward1964: 23; Maitland, Reference Maitland1908; Petit, Reference Petit2023).
Procedures developed in Westminster were transposed, initially imperfectly, to colonial governments in British North America. In the 1850s, when responsible government was in its infancy, a report on public accounts noted “…the Legislative Assembly of Canada often hears, for the first time, of large expenditures, many months after the cash is paid away” (Ward, Reference Ward1964: 31). Colonial governments were more intent on “nation building” than on structured expenditure controls. As one study of the period observes: “Economic development was a major goal of parliamentary procedure in the pre-Confederation period”Footnote 3 (O’Brien, Reference O’Brien1988: 413).
The Fenian Raids of 1866 led to the development of special warrants to solve a practical problem. By this time, “the system for ensuring parliamentary supremacy [over financial matters] was almost complete, on paper” (Ward, Reference Ward1964: 37). No sooner had a system been established than it was challenged by violent incursions of Irish nationalists from the USA.
“Under the former system,” writes H. R. Balls, “expenditures had been made without the sanction of Parliament; under the new [system], when the nation’s security called for action, expenditures were made in defiance of the directions of the Assembly” (Balls, Reference Balls1963: 186). As Norman Ward wrote in his foundational study of public finance in Canada:
…[In] 1866 the ministry found itself obliged, while the legislature was not sitting, to spend large sums for the defence of the province against anticipated Fenian raids which had not been foreseen in time to ask the legislature for sufficient money to pay the necessary bills. A violation of the law was the only possible alternative to a parliamentary grant, for the legislation of 1864 provided only for such unforeseen expenditures as repairs to public buildings…In due course, though not before Confederation, the scope of unforeseen expenditures that could be made on Governor General’s warrants was widened to include, as it still does, not merely repairs to damaged buildings, but “any expenditure not foreseen or provided for by Parliament [which] is urgently or immediately required for the public good (Ward, Reference Ward1964: 37).
The practical solution to the problem of urgently needed but unauthorized public expenditure was addressed more fully after confederation. Initially, this was through acts of indemnity for such expenditures and then, in 1871, by legislation providing for special warrants (Balls, Reference Balls1963: 187–8). In his study of special warrants, H. R. Balls describes early instances of their use federally, as well as changes in legislation. In 1931, the Consolidated Revenue and Audit Act (CRAA) was extensively amended, but “[the] special warrant provision was reenacted virtually unchanged” (Balls, Reference Balls1963: 190). The same was true when, in 1951, the CRAA was replaced by the Financial Administration Act (FAA). In 1958, the relevant provision was amended “to provide that any special warrant should be deemed included in, and not be in addition to, the amounts appropriated in the next Appropriation Act” (Balls, Reference Balls1963: 192). The situation then remained unchanged for nearly four decades.
After the November 1988 election, Parliament was convened, but approval of the estimates was deferred until September 1989. When parliamentary authorization for expenditures ended with the fiscal year, the executive repeatedly used special warrants to fund the operations of government. Peter Milliken, then an opposition member of parliament (MP) and later the speaker of the Commons, introduced a private members bill in 1989 to amend the FAA to restrict the use of special warrants. Even though Bill C-211 failed to pass in 1989, in 1997 he introduced a similar bill that was this time adopted. This legislative amendment was in direct response to the 1988–1989 controversy (Milliken, Reference Milliken1990). The “Milliken amendments” constrained the use of special warrants to two situations: (a) when Parliament is dissolved for a general election and (b) up to 60 days following the return of the writs following a general election (Milliken, Reference Milliken1990; Rasmussen, Reference Rasmussen1989; Government of Canada, 2007; Bosc & Gagnon, Reference Bosc and Gagnon2017).
Reforms to the federal system for special warrants have purposely limited executive flexibility by trying to anticipate and narrow when this instrument can be used. In March 2020, Parliament was adjourned due to restrictions arising from coronavirus disease 2019 (COVID-19); at the same time, major unforeseen expenditures were urgently needed for emergency health and economic measures. The use case for special warrants did not fit the narrow conditions prescribed by the Act. As a consequence, on March 13, Bill C-12, an Act to amend the Financial Administration Act (special warrant) was approved in a single day.Footnote 4 Bill C-12 temporarily relaxed the restrictions on the use of special warrants provided for by the Milliken amendments of 1997. There are parallels with the situation that faced government in 1866 in response to the Fenian Raids. The flexibility originally enabled by the innovation of special warrants is now severely curtailed in federal legislation.
Comparable constraints do not exist in provincial legislation (Appendix). Though relevant provincial statutes differ from the federal regime, they are similar to each other; there are, however, notable differences between provinces, including time constraints, conditions for spending and reporting requirements. These provisions are contained in the provincial FAA or similar legislation by another name.Footnote 5 All these statutes provide for special warrants to be used for unforeseen expenditures where it is needed “for the public good,” even when the legislature is neither prorogued nor dissolved. Special warrants cannot be used when the legislature is in session. For the legislature to be considered “not in session,” the duration of an adjournment ranges from 5 days in Nova Scotia to 30 days in Quebec. As provincial legislatures sit on average 50 days a year, this provides considerable scope for the use of special warrants (Gray and Cardoso, Reference Gray and Cardoso2019).
The purposes for which special warrants can be used vary: in Newfoundland and Labrador for “special reasons”; in Nova Scotia when it is “urgently and immediately required for the public good”; in Alberta when a minister attests to the urgency of the expenditure; and in British Columbia only in “states of emergency,” as defined in the Emergency Program Act. In Prince Edward Island, Ontario and Manitoba, there is no reference to “urgency” or “emergency.” Transparency in reporting also varies. In Newfoundland and Labrador, the Minute of Cabinet authorizing the special warrant is tabled in the legislature, but this does not set out the rationale for its use. In Ontario, a report explaining the rationale is provided to the treasury board and, thus, within the executive rather than to the legislature. Typically, special warrants are published in the provincial gazette; this records the use of the special warrant but not its rationale.
What is common to all is post facto approval by the legislature in supplementary appropriations. Before special warrants were developed, when faced with an urgent need for expenditures, the executive’s only option was to act without legislative authorization, as in the case of the Fenian Raids. Bateman views this type of decision as “unavoidable” rather than “sinister” (Bateman, Reference Bateman2020: 145). To regularize the situation, the legislature would subsequently adopt an indemnifying bill to exonerate ministers and public servants (Ward, Reference Ward1964: 37). Statutory provision for special warrants replaced post facto “indemnification.” “Viewed with maximum optimism,” writes Bateman, “the fact that parliaments retrospectively appropriate unlawful expenditure evidences their authority to censure executives that behave unlawfully.” Viewed less optimistically, retroactive approvals reveal “the hard limits of parliaments’ financial authority” (Bateman, Reference Bateman2020: 145).
Conventions on the Use of Special Warrants
Restraint on the use of special warrants is a constitutional convention. Conventions are unwritten norms that solve elite coordination and/or cooperation problems over the exercise of political power (Barry et al., Reference Barry2025). The difference between constitutional conventions and mere political conventions is that the latter—for example, the alternation between English and French leaders of the Liberal party of Canada—only controls political actors, whereas the former—for example, resignation of the ministry on loss of confidence of the legislature—controls political actors use of state power.Footnote 6 Moreover, not all constitutional conventions exhibit the same level of restraint. The convention related to the use of special warrants is closest to a constitutional “semi-convention,” which can be occasionally disregarded without significantly undermining the constitutional principle of parliamentary control over public finance (Heard, Reference Heard1989: 73).
Conventions on the use of special warrants are constitutional because they restrain the use of the Crown’s powers (Smith, Reference Smith1995: 18). This is supported by precedent. From the time of their inception until the late twentieth century, the use of special warrants was minimal and infrequent.Footnote 7 In 1949, R. MacGregor Dawson described the use of special warrants as “severely restricted” even though, at that time, the CRAA provided a relatively wide scope for their use (Dawson, Reference Dawson1949: 430). Dawson’s comments reflect his view as to convention rather than the loose legal constraints. Their treatment by J. R. Mallory is also instructive, especially given his concern with the “aggrandizement of the executive” (Mallory, Reference Mallory1957: 113).Footnote 8 In his 1971 The Structure of Canadian Government, Mallory refers to special warrants as a “perfectly constitutional device”; however, he notes that they are “rarely used,” in part because of embarrassment from “a confession of miscalculation or failure” (Mallory, Reference Mallory1971: 136). Restraint on the use of special warrants is referred to by D. Michael Jackson as a “parliamentary convention,” which if flaunted would be “borderline unconstitutional behaviour” (Jackson, Reference Jackson and Leeson2008: 18).
How legislatures, academics and the public have responded to perceived contraventions of the constitutional convention that restrains the use of special warrants is illustrated by three controversies over the past century, two federally under prime ministers Diefenbaker and Mulroney and the third in Saskatchewan under Premier Grant Devine. The first, involving the Diefenbaker ministry in 1957–1958, was not so much controversial for the purpose (funding to settle Hungarian refugees) or the timing (during dissolution for a general election) of the special warrant, but rather owing to an opposition demand that Parliament should review such expenditures post facto (Sager, Reference Sager1961; Mallory, Reference Mallory1971: 136–7). Donald Fleming, minister of finance, argued in reply that all that was needed to satisfy the legal requirements of the FAA as well as any constitutional convention was “deemed appropriation” by inclusion of the expenditures in a subsequent grant of supply. Advocating the existence of a convention, the opposition relied on the experiences of two members of parliament who had been provincial finance ministers as well as earlier committee testimony from the deputy minister of finance, all of whom favoured parliamentary review and retroactive parliamentary approval (Sager, Reference Sager1961: 312). In 1958, the FAA was amended, making clear that special warrants are subject to post hoc parliamentary review, reflecting not only the principle of lawful spending but also legislative scrutiny of expenditures. This continues to be the case, although as we see below, some provincial jurisdictions are diverging in practice.
Controversy in Saskatchewan over special warrants arose three times, in 1982–1983, 1986–1987 and 1991. The government of Grant Devine, on two occasions, used special warrants after a general election to spend beyond the end of the fiscal year; subsequently, and even more controversially, it also used special warrants to finance all public services from the beginning of the fiscal year in 1991 through until late September, when there was a general election (Massie, Reference Massie2020: 211–6; Jackson, Reference Jackson2013: 152–6).Footnote 9 In a highly critical article dealing with the Mulroney and the Devine post-election contraventions, Merrilee Rasmussen—who was also the author of a subsequent legal opinion that became public during the controversy in Saskatchewan in 1991 (Rasmussen, Reference Rasmussen1991)—stated that these “governments have demonstrated a narrow approach to understanding statutes that ignores the relationship between convention and law” (Rasmussen, Reference Rasmussen1989: 17).
Although sanctions for breach of convention are more often political than legal, that does not make them any less effective. In the Diefenbaker controversy, the opposition gained acceptance of a requirement for Parliamentary review of special warrants, which was subsequently enacted in the FAA by the Diefenbaker government. In the Mulroney controversy, the Senate chided “a virtually unprecedented move” that, although within the letter of the law, “clearly went against Parliamentary practice and should not be used as a precedent by any other administration” (Milliken, Reference Milliken1990: 24). In the Devine controversies, the use of special warrants placed two consecutive lieutenant governors in awkward positions by making their decisions appear overtly political and incited legislative debate over the existence and nature of the constitutional convention (Massie, Reference Massie2020: 212–5; Jackson, Reference Jackson2013: 153–4). Recent practices in the provinces, though also inconsistent with convention, have engendered comparatively less scrutiny or controversy. They have been framed as rule bending rather than breaking, as befits a constitutional semi-convention (Heard, Reference Heard1989; Miragliotta and Algra-Maschio Reference Miragliotta, Algra-Maschio and Barry2025).
Practice of Spending by Special Warrant
In February 2024, news media reported on the use of special warrants in Prince Edward Island, Manitoba and Saskatchewan (Brun, Reference Brun2024; Winnipeg Sun, Reference Sun2024; Salloum, Reference Salloum2024).Footnote 10 What attracted media attention was that special warrants were used as a routine, rather than as an extraordinary, measure. The relative amounts involved are noteworthy. Expressed as a dollar amount and as a percentage of total expenditures for 2023–2024, special warrants were used for $757 million in Saskatchewan (3%), $710 million in Manitoba (3%) and $327 million (10%) in Prince Edward Island. By comparison, the federal government used special warrants for expenditures of $18.5 million or 0.03 per cent of total expenditures in 1939–1940, at the start of the Second World War.
Although framed by the media as irregular, these instances, perhaps counterintuitively, actually typify a pattern of increased use of special warrants over recent decades by provincial governments for routine spending. In 1996, the auditor general of British Columbia (BC-AG) noted that “[n]ineteenth century conditions—great distances, slow transportation, difficult communications and parliamentarians meeting for only a small portion of the year” could justify the use of special warrants (BC-AG, 1996).Footnote 11 The BC-AG went on to quote a recommendation by a 1981 provincial task force on public finances to “eliminate the present routine use of special warrants for expenditures which cannot truly be described as ‘urgently and immediately required for the public good,’ and would thus reserve this instrument for unforeseen emergencies” (BC-AG, 1996: 7). The BC-AG also noted that, in the early 1990s, leaders of the provincial New Democratic and Liberal parties criticized the use of special warrants. “However, despite all this public commentary dating back more than 15 years,” wrote the BC-AG, “…the use of special warrants to authorize government spending has increased” (BC-AG, 1996: 14). British Colombia is not exceptional in this practice.
The auditor general of New Brunswick (NB-AG) noted that since 2011 the government has regularly overspent without legislative approval and that post facto legislative authorization had not been obtained in a timely manner (NB-AG, 2018: 134).Footnote 12 In 2024, a report by the auditor general of Prince Edward Island (PEI-AG) dealt with both the increase in the amount of expenditures authorized by special warrants and the timing of expenditure and retroactive approval (PEI-AG, 2024: 99). In such instances, legal boundaries are stretched and convention discarded, undermining the legislature’s role (Campbell and Brun, Reference Campbell and Brun2024). Recent records of provincial spending by special warrant as well as reports from auditors general in multiple provinces spanning decades illustrate the partial erosion of a constitutional convention that once restrained executive spending by special warrant. This is demonstrated across jurisdictions, by various political parties and over four decades.
Discussion
Special warrants were developed to deal with situations where expenditures, when urgently needed, had not been authorized by the legislature and where obtaining legislative authorization in a timely way was not practically possible. Special warrants provide a lawful means to deal with excess spending in emergencies. However, the statutory basis for the use of special warrants does not fully incorporate a requirement for exigent circumstances. This leaves to convention interpreting the appropriateness of use. What historical analysis of spending by special warrant reveals is not so much the breaking but the bending of convention.
Controversies during the Diefenbaker and Mulroney ministries led to legislative amendments to, firstly, secure in federal law the right of Parliament to retroactively review and approve spending by special warrant and, secondly, to restrict the terms of use. There has been no “diffusion of innovation” for comparable legal reforms in the provinces (Poel, Reference Poel1976). This makes possible the routine use of special warrants as a convenient legal “grey hole.” Notwithstanding intermittent criticism by the provincial media or auditors general, rarely does elite or mass public concern rise to the level of controversy seen in the Devine case in Saskatchewan. This is so despite a documented pattern of provinces using special warrants in recent decades as a convenient way to exceed approved annual appropriations.
This raises research questions about the way in which constitutional conventions change, the legitimacy of change and the way we know when change has occurred. Where practice no longer conforms to convention, it is thought that the convention is modified or ceased; change is possible, but alteration cannot be so fluid and frequent as to merely become a guideline (Heard, Reference Heard2012). It is perhaps true that, in some circumstances, constitutional conventions are changed by “stealth” (Albert, Reference Albert2015), as a result of “populism” (Barry, Reference Barry and Barry2025; Mailey, Reference Mailey2019; Walker, Reference Walker2019), or in a preference for written rules rather than discursive interpretation by elites in accordance with the circumstances (Aucoin, Reference Aucoin2011). The case of special warrants is satisfied by none of these explanations. When the constitutional convention on the use of special warrants is contravened so severely as to undermine the constitutional principle upon which the convention is based, as it was with Devine in 1991, a corresponding level of political sanction follows. This is not so when the contravention is minimal or temporary or conditional, as with recent provincial cases, and where political actors still profess to be generally restrained by convention in broad adherence to the constitutional principle. Arguably, it has become reasonable for finance officials to advise ministers that using special warrants for routine excess spending, and not just for emergencies, is consistent with past practice, even if contrary to constitutional convention. In light of practice over the past 40 years, perhaps we need a theory of change in constitutional conventions that better reflects empirical data on the pace, volume and incrementalism of government decision making (Scheuerman, Reference Scheuerman2004; for example, Campagnolo, Reference Campagnolo2022).
More recent scholarship offers an alternative to the typical binary view of respect for constitutional conventions, whereby it is possible for actors to claim a temporary or partial exemption to adherence to a constitutional convention with respect to what is permissible, to whom the convention attaches, and when it must be observed (Miragliotta and Algra-Maschio, Reference Miragliotta, Algra-Maschio and Barry2025). Without regard to aberration, others will be expected to follow the convention in the future. Special warrants seem to fit this case, where the convention has not fallen into complete disuse or fundamental alteration, and where outright contravention is still rejected by opposition parties, elites and the public, but limited claims to partial exemption are no longer forbidden. This sets up a future line of research for students of Canada’s unwritten constitution about where and how exemptions facilitate a convention’s evolution.
This research note also raises questions about the separation of powers in a fast-paced society. One might think the discursive (and thus slower) nature of legislatures is a dead letter in an age of accelerating government decision making. The decline of Parliament has been bemoaned by political scientists for generations, though there is limited empirical study on the practical influence of parliamentary procedure on the exercise of power (Brodie, Reference Brodie2018: 78–79). Even though parliamentary scrutiny of spending is foundational to Westminster democracy, a lot has changed since the Fenian Raids (Good, Reference Good2011; Reference Good2017; Goldenberg, Reference Goldenberg2006; Maslove and Cutt, Reference Maslove and Cutt1989). Given that Canada’s Westminster parliamentary democracy is lauded for its adaptability (Smith, Reference Smith2017; Scott, Reference Scott1945), scholarship should account for this fundamental characteristic. We now broadly accept other between-budget innovations such as fall economic outlooks (“mini-budgets”) and formal administrative mechanisms for “off-cycle” spending proposals. Perhaps scholarly descriptions of the annualized budget cycle of review and approval insufficiently reflect contemporary rhythms of public finance. We need research that documents how the machinery of government works in Canada, based not only on the federal government but also provincial practices; this should be based not just on normative claims about how it ought to be but also on patterns of change from historical analysis.
Appendix: Compendium of Statutory Provisions for Special Warrants
Open access
