From: John Lester
To: Members of the Treasury Board
Date: April 8, 2024
Re: The federal expenditure management system needs a makeover
The federal expenditure management system looks good on paper.
Transparency is served by publication of five-year spending plans for major spending categories in the annual budget and detailed information in the main estimates and departmental plans.
Efficiency and effectiveness are served by setting objectives for program spending and requiring departments to report on the achievement of these objectives. This result-based management framework is buttressed by requiring most spending programs to be evaluated on a five-year cycle. Transparency is further served by making these reports publicly available.
Dig a little deeper, however, and flaws appear.
The first is incomplete coverage of spending. Governments use the tax system to achieve economic and social objectives, making these initiatives functionally equivalent to spending programs – think of the tax incentives given to electric-vehicle battery manufacturers and the Goods and Services Tax credit available to Canadians with low incomes, particularly when used to deliver a grocery rebate. I estimate that tax-based spending programs will cost about $60 billion in 2024, but these programs are not included in the government’s expenditure plan and there is no requirement for them to be evaluated.
The second flaw is that program objectives are poorly specified, so that program effectiveness is not properly assessed. For example, business subsidy programs are usually assessed by considering the responses of subsidized firms: Did investment, output, and the number of jobs increase? But these aren’t much of a test to pass – it would be flabbergasting if firms did not respond favourably to a subsidy. Evaluators can always point to some job creation, or at least job protection, arising from a subsidy.
The real question is whether Canadians overall are richer or poorer because of the business subsidy. In a labour-short economy, subsidies simply move workers from one industry to another. This shift will only raise real incomes if workers are more productive in the subsidized industry than in another industry by a margin large enough to offset the costs of the subsidy. These include expenses required to administer the programs, the costs incurred by firms to comply with the program requirements and the economic damage caused by funding initiatives with higher taxes. Determining the impact of a business subsidy on real income requires a rigorous analysis of these costs and benefits, but this is not standard practice in program evaluations.
A third flaw is that spending programs – tax and non-tax – are introduced with at best a superficial assessment to determine if there is a reasonable expectation they will provide value to Canadians. This is in marked contrast to regulatory initiatives, which must pass a benefit-cost test before they can be implemented. The federal government recently committed $18.8 billion in subsidies to Stellantis NV, LG Energy Solution Ltd. and Volkswagen AG to produce EV batteries, without making a solid case that Canadians will be richer as a result.
The subsidy for Volkswagen was sold as paying for itself over five years, apparently through tax revenues generated directly and indirectly by the project. This is problematic because there are no details on the calculation and because even the best break-even calculation is not a reasonable basis for a decision. A basic point is that workers in the battery plants would have been employed elsewhere, so a break-even analysis is inherently misleading. While it may not be possible to perform a complete analysis of benefits and costs prior to offering the subsidy, with the huge amount of money at stake, a preliminary analysis should have been prepared and shared with Canadians who are footing the bill.
The fourth and most significant flaw is that the government has no budget constraint. While five-year spending plans are announced each year in the budget, they can be changed on a whim. Since the COVID-19 pandemic began in 2020, program spending has been increased in successive budgets and economic and fiscal updates such that program spending in the current fiscal year will be almost $60 billion, or 0.4 percent of gross domestic product, higher than projected in the 2020 economic and fiscal update. Changes caused by a cyclical downturn are to be expected, but governments should set out a plan for non-cyclical spending at the beginning of their mandates and stick to it over that mandate. For example, the government could commit to keeping non-cyclical spending below a certain percentage of cyclically adjusted GDP.
Without a cap on discretionary spending, governments have no incentive to fund new initiatives by eliminating programs that are ineffective or no longer consistent with current priorities. With a cap, the $7 billion spending in 2023-24 on new programs to promote a clean economy would have squeezed out some spending on legacy programs, instead of adding to overall spending.
Even without considering pandemic-related initiatives, growth in program spending under the Liberals has been reckless. A revamped expenditure management system would inject some needed discipline into the process.
John Lester is a fellow-in-residence at the C.D. Howe Institute and an executive fellow at the University of Calgary’s School of Public Policy.
To send a comment or leave feedback, email us at blog@cdhowe.org.
The views expressed here are those of the author. The C.D. Howe Institute does not take corporate positions on policy matters.
A version of this Memo first appeared in The Globe and Mail
