Auditor General Sheila Fraser’s quarterly report, released this week, had two items of concern to advisors: whether trusts — not income trusts, but inter vivos and testamentary trusts — are properly reporting their income, and whether the CRA has kept track of RSP overcontributions.

While Canada’s tax system is based on self-assessment — individuals report their income voluntarily and pay the taxes based on that reported income — CRA has a number of risk management systems that spot errors, both inadvertent and deliberate, to gather most of the taxes that individuals legitimately owe.

Fraser’s office found that tax errors have increased from less than 6% of all returns filed by Canada’s 24 million taxpayers in 1997 to 11%, as of 2003, accounting for a potential tax loss of $586 million. CRA, and its predecessors, recovered only 27% of that amount, $160 million, and that came from reassessing just 3% of tax returns.

How does CRA spot tax errors? First of all, it checks the deductions and credits a taxpayer can claim against the permitted threshold amounts. That data is also fed into a scoring system, based on all 32 credits and deductions, which determines the potential tax at risk and which results in about 1.5% of tax returns being selected for review — which may lead to a taxpayer being reassessed for taxes owing. CRA then selects 2.7% of tax returns for review — not just for tax errors, but also to test the statistical program CRA uses against returns that are compliant. By using automated systems rather than picking returns at random for review, the auditor general’s office says CRA generally recovered four times as much revenue.

Finally, it matches individual taxpayer claims against those submitted by third parties: T-4s and T-5s provided by employers, financial institutions, stock brokers. Where there are significant discrepancies between the individual’s reporting and third-party reports, the tax return is flagged for the amount of income tax that could be recovered. Based on these three programs, CRA selects returns for auditing.

However, in the absence of detailed third-party records for trusts, or simply by failing to match those records, the auditor general’s office suggest that there is a significant tax loss, even though CRA does audit a proportion of trust returns. As a result, the auditor general’s office implies, CRA doesn’t have a full handle on forgone taxes, nor does it have a system that would flag the trusts with the most potential for tax recovery.

It’s worse with RRSP overcontributions, which are limited to $2,000, and attract a 1% a month penalty for amounts above that threshold. Although CRA could check taxpayers’ RRSP holdings against their brokerage accounts, it doesn’t. Instead it relies on taxpayer self-reporting. Still, the CRA has said that, for the 2004 tax year, it will start comparing what individuals reported as their RRSP holdings against information financial institutions have provided to the tax agency.

The auditor general’s offices also says CRA should make better use of its matching process so that income attributions and losses reported by partnerships, flow-through-share companies and tax-shelter promoters can be used to verify taxpayer claims for losses and credits.

It’s trusts that Fraser’s office emphasizes. There were 175,000 in the 2003-2004 audit year, and they paid out $23 billion to beneficiaries, and $3 billion in federal, provincial and territorial income taxes. While T-3 slips report the amounts received by each beneficiary, what gets reported to the beneficiary is only the total disbursement, not the proceeds of each individual trust. As a result, it’s not clear that what the trust reports as a deduction for income flowed through to beneficiaries is actually taxed in the beneficiaries’ hands.

Fraser’s office also says that CRA does not systematically figure the potential tax loss. In particular, inter vivos trusts came under the auditor general’s spotlight, since they are less likely to be subject to audit. An inter vivos trust allows someone, for example a parent, to retain control over the assets, while distributing the income from the trust to other beneficiaries — e.g., the children — in whose hands the income is taxable. A testamentary trust, by contrast, is one established through the death of the owner of the assets, for the benefit of heirs. It has to pay tax on capital gains and retained income every 21 years.

CRA reovered $22,300 in extra taxes on average for every inter vivos trust it audited for non-compliance, compared to $4,500 for every testamentary trust audited. With testamentary trusts, the auditor general’s office says that CRA is not following up to ensure that the information slips trusts send in are complete and accurate. It selected 20 trust tax returns from a population of 451 that would fall under the 21-year rule, and found that CRA should have verified the information for at least eight of them, but did so only for three.

“Such a high error rate shows that controls can be improved,” Fraser’s report said. Its conclusion: better estimates of the taxes at risk for domestic trusts that CRA selects to verify and a requirement that trusts submit a statement of assets and liabilities along with their income tax returns.

CRA says it is reviewing the situation and the cost-benefit analysis of having trusts submit balance sheets along with their tax returns.

Filed by Scot Blythe, Advisor.ca, scot.blythe@advisor.rogers.com.

(11/24/04)