Munir Sheikh is former chief statistician of Canada and former federal associate deputy finance minister. He is author of the Macdonald-Laurier Institute publication Estimating the True Size of Government: Adjusting for Tax Expenditures.
"Boutique" tax measures to target and assist certain groups of Canadians have created considerable concern recently because they favour a small number of taxpayers at the expense of the general population. These tax expenditures raise issues of fairness, and they make an already complex tax system even more complex.
In recognition of this, the Liberal government's first budget committed to undertaking "a review of the tax system to determine whether it works well for Canadians, with a view to eliminating poorly targeted and inefficient tax measures."
This review is welcome, as it should contribute to improving the tax system, considering the value of federal tax expenditures is likely almost half of direct expenditures. Some of the largest tax expenditures relate to retirement savings, taxation of capital income, charitable donations, small-business taxation and the GST/HST credit.
However, in my view, given the opportunity, the government should go further with its review and deal with the core of the problem: namely, the incentives that currently exist for governments of all stripes to use tax expenditures with little scrutiny. Even if this review improves the tax system, the existence of underlying incentives to overuse tax expenditures means the almost certain deterioration of the tax code over time.
When used properly, tax expenditures can provide the biggest payoff at least cost for a range of economic and social policy goals. For example, to help Canadians generate retirement income, the tax system provides deductions for contributions to registered retirement savings plans. In this case, the use of a tax expenditure is optimal because tax assistance is tied to a taxpayer's income level.
However, even that very positive example demonstrates an important problem with the use of tax expenditures.
First, a government's decision to provide increased retirement income support indicates an increase in the size and role of government. But under current accounting, this increased tax support would actually show up as a reduction in government revenue, suggesting the opposite – that the role of government had declined. Second, the information on the cost of tax expenditures is not transparent, as it is buried in the resulting lower tax revenue. Third, this non-transparency makes tax expenditures attractive to politicians. Those on the left use the tax system to effectively increase spending, while those on the right think of it as reducing taxes.
In a paper I wrote for the Macdonald-Laurier Institute, I showed that adding total tax expenditures to either total government expenditure or to taxes reverses the declines in the expenditures-GDP and revenue-GDP ratios that can be otherwise noticed in the recent past. So government was in fact growing, rather than shrinking. This in itself is neither good nor bad. As noted earlier, tax expenditures can be a useful policy tool – what must change is the incentives inherent in the current system that lead to overuse and misuse.
Direct expenditures for government programs, and tax expenditures that achieve government objectives by another means, should be put on a level playing field. I recommend four principles:
1. All proposals for tax expenditures should be evaluated by whether they would be acceptable to the government if they were to be introduced transparently as direct expenditure programs.
2. Having passed such a test, a proposal should be considered for delivery as tax expenditure only if it can be demonstrated that this mechanism would be administratively more efficient than a new expenditure program.
3. The proposed tax expenditure program should be fully costed.
4. All tax expenditures should be reported in the annual budget with their total cost side by side with, and added to, total direct expenditures.
The last suggestion needs some explanation. Now, the federal Finance Department puts out an annual report providing individual estimates of lost revenue for more than 200 tax expenditures. The department cautions against aggregating the information and provides good reasons. However, this results in tax expenditures being less transparent and, therefore, open to misuse.
Given the importance of this issue, I have argued that we should follow the example of the Organization for Economic Co-operation and Development (OECD) and aggregate these data by dealing with the estimation challenges, rather than by avoiding them. In my view, it is possible to produce quite reasonable estimates.
By making tax expenditures transparent in annual budgets, the incentive for politicians to overuse them would be diminished. It is hoped, then, they would be used only when they are the most appropriate policy tools. This should improve accountability and lead to better outcomes for Canadians.