This Web page has been archived on the Web.

2005 November Report of the Auditor General of Canada

Main Points

What we examined

The Canada Revenue Agency is responsible for administering the Income Tax Act. It strives to ensure that Canadians pay their required share of taxes and that the tax base is protected. We looked at how the Agency verifies the accuracy of income tax returns of domestic trusts and certain aspects of personal income tax returns. We examined how the Agency decides which tax returns it will verify; whether it has adequate verification procedures, including its use of information from third parties; and whether it compiles complete and accurate information on the results of its verification programs.

Why it's important

The Canadian tax system is based on self-assessment, with each individual and trust required to provide complete and accurate information to the government on the income taxes they owe.

The programs we looked at form the backbone of the Agency's verification of personal income tax returns. In 2003–04, these programs scrutinized about 24 million returns filed by taxpayers, and they verified aspects of returns that did not appear to comply with the law. Personal or individual income tax is the single most important source of government revenue; in 2003–04, the Agency assessed $125 billion in taxes payable by individuals to the federal, provincial (other than Quebec), and territorial governments. In the same year, domestic trusts were assessed $3 billion in federal, provincial, and territorial income taxes, and they allocated income of $23 billion to their beneficiaries.

What we found

  • The Agency's processing review program has a well-designed and well-executed risk-based approach for selecting and verifying deductions and credits that individuals have claimed on their tax returns but may not be fully entitled to. It found that the percentage of such taxpayers doubled to 11 percent from 1997 to 2003 although according to our estimate, the total revenue at risk has remained relatively constant for the last four years. Using the Agency's data for 2002–03, we estimated that, had all taxpayers complied fully with the rules for claiming the 32 deductions and credits covered by the processing review program, revenues from personal income taxes that year would have been $586 million higher. The program recovered 27 percent of that amount, or $160 million, by verifying about three percent of the returns assessed that year.
  • In its matching program, the Agency compares the information reported by taxpayers with information submitted by third parties such as employers and financial institutions. Where it finds one or more discrepancies, it calculates the amount of income tax that is potentially recoverable. While the matching program selects returns for review based on risk, its estimate of risk does not include the system's calculation of the amount potentially recoverable. As a result, the Agency does not consistently select and review the returns with the largest amounts of potentially unreported income. Nor does it estimate the tax at risk for returns where it has identified discrepancies but not verified them.
  • In not monitoring the tax revenue impact of non-compliance in the areas verified by the processing review and matching programs, the Agency is not considering an important element in evaluating these programs' effectiveness.
  • The Agency does not systematically evaluate the tax revenue at risk in domestic trusts when choosing the tax returns it verifies. In addition, the development of an effective risk evaluation system is hampered by the lack of key information such as the value of assets and liabilities held in trusts. The Agency emphasizes audits of testamentary trusts. In the last three years, average tax recoveries from audits of testamentary trusts have been about five times smaller than average recoveries from audits of other trusts.
  • There are also deficiencies in the Agency's review activities for tax returns of domestic trusts. For example, the only measure of its performance in this area is whether a return was processed in the time allotted by the Agency's service standard; it lacks information on corrections made by its assessors; and it does not compare deductions claimed by trusts for allocations to beneficiaries against the amounts reported on the information slips the trusts provide to beneficiaries.

The Agency has responded. In its response to each recommendation, throughout the chapter, the Canada Revenue Agency has indicated the action it has taken or plans to take.

Introduction

3.1 The Canada Revenue Agency received about 24 million income tax returns from individuals in 2003–04. This represents the single most important source of government revenue, with taxes of $125 billion assessed that year. The Canadian tax system is based on self-assessment, with each individual required to file a tax return that reports complete and accurate information and to pay taxes he or she owes under the Income Tax Act. Non-compliance can result from deliberate acts of tax evasion as well as from non-deliberate honest mistakes in interpretation and mathematical miscalculations.

3.2 The Agency's verification programs for individual or personal income tax returns include three validation programs—confidence validity, processing review, and matching—and four desk audit programs—refund examination, post review, office audit, and office audit of taxpayer-requested adjustments. The three validation programs form the backbone of how the Agency verifies personal income tax returns. These three programs scrutinize every one of the returns filed by taxpayers every year and verify aspects of returns that do not appear to comply with the law.

3.3 The Income Tax Act also requires trusts to file tax returns. According to Agency data, in 2003–04 about 175,000 domestic trusts reported that they had allocated income of $23 billion to beneficiaries, and that they themselves were assessed $3 billion in federal, provincial, and territorial income taxes. The Agency's verification programs for domestic trusts include a compliance review when the return is initially processed and audits of selected returns.

Focus of the audit

3.4 The objective of our audit was to assess how well the Agency identifies and manages the risk that individuals and domestic trusts might fail to comply with the law when preparing their income tax returns, and how well it uses information from third parties in verifying personal income tax returns.

3.5 We looked at whether the Agency

  • ranks tax returns for verification by level of risk,
  • has adequate procedures for verifying all the information in a tax return,
  • makes reasonable use of information slips that the law requires third parties to submit,
  • accurately and completely captures results of its verification activities, and
  • addresses known areas of non-compliance.

3.6 We did not look at the Agency's programs for verifying compliance with the rules that govern the reporting of income from self employment, as we had looked at them last year. Nor did we examine specialty audit programs, such as tax avoidance and international tax, or the Agency's compliance activities for non-resident or foreign trusts. Further details on our audit objective, scope, approach, and criteria are presented in About the Audit at the end of the Chapter.

Observations and Recommendations

Verifying personal income tax returns

3.7 To ensure that all its verification programs for personal income tax returns use their resources as effectively as possible, the Agency identifies areas where taxpayers are more likely to have difficulty completing their tax returns or to not comply with the law. It then identifies specific tax returns with higher potential for tax recovery. A number of automated risk assessment tools screen all returns (Exhibit 3.1), to determine the likelihood that a return contains incomplete or inaccurate information and to estimate the additional income taxes owed, if the suspected irregularities were confirmed and corrected.

3.8 Even before applying risk assessment criteria, the Agency subjects every tax return to an initial review to ensure that the return is consistent with information already on file, that calculations are correct, that certain deductions and credits claimed by the taxpayer do not exceed the maximum amounts allowable, and that the taxpayer is entitled to the amounts he or she has claimed.

3.9 Then a scoring system assigns a score to each of 32 specific deductions and credits in a return. Based on those scores, the confidence validity program verifies selected returns. This program also verifies returns that the regular screening system has selected because they contain losses, deductions, or credits that exceed established amounts. In total, the confidence validity program selects about 1.5 percent of personal tax returns for review.

3.10 At this point, the Agency sends the taxpayer a notice of assessment that confirms or changes the taxpayer's calculation of taxes payable and any refund due.

3.11 Once the notice of assessment is sent, the processing review program selects about 2.7 percent of personal income tax returns for review. The program selects returns mainly by their assigned score but also selects some at random to test and develop the Agency's risk assessment criteria.

3.12 The Agency also matches certain information on the tax return to selected information provided by third parties—such as remuneration paid, as shown on the T4 slips provided by employers—and information from the tax returns of spouses and common-law partners. The matching system identifies all returns with discrepancies, and the matching program verifies a number of those returns.

3.13 The refund examination, post review, and office audit programs also select returns for desk audit. Each of those programs has a defined target. For example, as part of their workload

  • the refund examination program verifies the return if a taxpayer asks for a refund that exceeds a set amount,
  • the post review program verifies claims for "other employment expenses," and
  • the office audit program verifies taxable capital gains.

These three programs select returns for audit based on the estimated amount of tax revenue at risk.

3.14 When any of the Agency's verification programs for personal income tax returns select a return for review, examiners will normally verify only the claims or income items identified at risk, not all the amounts on the tax return. If the review results in changes to the return, the Agency sends the taxpayer a notice of reassessment.

Verifying deductions and credits

The Agency has a well-designed and well-executed approach to selecting income tax returns for audit

3.15 The confidence validity program automatically selects returns for review according to predetermined criteria known as confidence validities. These are designed to evaluate the level of risk associated with specific items or claims on a tax return. "Regular" confidence validities are based on simple logic and on whether the amount of the claim exceeds predetermined thresholds. The regular validities apply to certain deductions, credits, and losses. "Scored" confidence validities are based on how the scoring system evaluates the level of risk associated with 32 different deductions and credits that can be claimed on a tax return and on whether the taxpayer is asking for a refund. A higher score indicates a higher risk of tax loss. Most deductions and credits that the system scores appear on page 3 and Schedule 1 of the personal income tax return (Exhibit 3.2).

3.16 The scoring system uses many criteria that subject matter experts develop. Through the interrelationship of these criteria, a score is assigned to specific deductions and credits in each return; a taxpayer's compliance history with the Agency also affects the score. Returns are sorted by categories with each of the 32 deductions and credits that the program verifies and each of the four ways or modes of filing (paper, EFILE, NETFILE, TELEFILE) representing a category. The processing review program then verifies a number of returns selected from each category; the returns with the highest scores are selected for verification. The staff resources available to do the verification and the estimated level of tax dollars at risk for each category, as compared to the other categories, determine the number of returns that are reviewed from each category.

3.17 Each year, a statistical analysis of the results of the processing review program allows the Agency to determine how effective the scoring system has been at detecting non-compliance. Part of this analysis includes comparing the number of returns changed or "adjusted" in each category with the actual number reviewed, and the average amount of tax recovered for each review. Tax officials also validate the scoring system, by comparing the tax recovered in selecting claims according to their assigned score, with the tax recovered in selecting claims at random. For 2002–03—the last year we reviewed where the Agency had completed this analysis—the scoring system was effective. Selecting claims by their risk score recovered more revenue than selecting claims at random—on average, four times more per review—in all but two categories of deductions and credits.

3.18 Some high-risk claims were not verified by the processing review program. Although the confidence validity program reviews tax returns that meet its criteria throughout the year, the processing review program did not verify returns filed after a certain date. Program officials estimated that in 2003–04, about 26,000 tax returns that met the program's selection criteria for verification were received after the cut-off date and therefore were not verified. The Agency made system changes to correct this deficiency in time to verify the 2004 income tax returns.

Verifying income subject to third-party reporting

The matching program identifies potentially unreported income

3.19 In the fall of each year, the Agency matches information on tax returns with information that employers, financial institutions, and other third parties provide. In addition, the tax returns of individuals who are married or living common law are compared with their spouses' or partners' returns. This comparison is designed to identify

  • unreported income;
  • an incorrect claim for an amount of "income tax deducted;"
  • credits and deductions that exceed the limits;
  • net family income for the purpose of the Canada Child Tax Benefit, the Goods and Services Tax Credit or the Harmonized Sale Tax Credit, and the Guaranteed Income Supplement; and
  • incorrect "pension adjustment" amounts.

3.20 When the matching system identifies a discrepancy, it calculates how much additional income tax would be owed if the discrepancy was corrected. All returns that could yield more than a predetermined amount in additional federal, provincial, or territorial income taxes or that could result in adjustments in the taxpayer's favour are candidates for review. These returns are grouped according to the source and nature of the discrepancy. Tax officials then verify a portion of these returns.

3.21 The reasons for selecting the returns verified by the matching program were not fully documented. The resources allocated to the matching program resulted in the review of 55 percent to 80 percent of the discrepancies identified by the system each year over the past five years. We were told that program officials use several criteria to decide which returns to review. The criteria include the source and nature of the discrepancy as well as how often changes were made when returns with similar discrepancies were reviewed in previous years and how much additional tax was recovered, on average, when a change was made. However, officials were not able to provide the full details to show why the matching program selected the returns it verified in 2004–05. In the absence of supporting documents, we were not able to determine whether the Agency followed its own criteria in selecting returns for verification that year.

A number of factors have made the matching program less effective

3.22 The matching program does not consistently select the riskier returns for verification. The categories that the Agency uses to select the returns that the matching program will verify are based on criteria that indicate the risk of non-compliance. However, these criteria do not include the amount that the system has already calculated as potential taxes that can be recovered on a return. As a result, some returns it verified had less tax potential than other returns it did not verify.

3.23 The Agency has enough information to calculate the expected tax recovery of most returns that contain discrepancies. The Agency could select returns for review, starting from those with the highest expected tax recovery. This could be done, no matter how else the returns have been categorized. Once selected according to the highest expected tax recovery, the returns could be then re-sorted by category to facilitate the workflow in the taxation centres. Returns in specific categories could still be selected and verified for purposes other than maximizing dollar recoveries—for example, to educate taxpayers on reporting a particular type of income or to gain information on how taxpayers report different types of income.

3.24 The matching program currently does not incorporate other readily available information that could be useful in assessing the risk that a taxpayer has incorrectly reported income. For example, information about the taxpayer's past compliance history and about whether the return has been selected for review by the confidence validity or processing review program is not included in the matching program criteria.

3.25 Matching T3 information slips has been a challenge. The law requires that an income tax return include income the taxpayer received from trusts. Trusts are generally required to report the income they allocate to beneficiaries on T3 information slips.

3.26 In reviewing program information and interviewing program officials, we found that the matching of T3 slips to returns has produced poor recovery of taxes owed because of difficulties with identifying and confirming discrepancies. For example, on the T3 slips provided to the Agency, financial institutions report each separate amount distributed to a beneficiary from each of the beneficiary's trusts. However, on information slips provided to each beneficiary, the institutions report one amount representing the combined payments from all the trusts.

3.27 The Agency's regular workload does not include verifying discrepancies between T3 slips and income tax returns and assessing the comparative risks of tax loss, if a taxpayer fails to report trust income or income from other sources. These verification activities are only conducted when the Agency specifically allocates funds for that purpose.

3.28 The matching program recovered lower revenues in a recent two-year period. Revenues recovered through the matching program declined significantly in 2002–03 and 2003–04 (Exhibit 3.3). In their 2002–03 report, program officials identified a number of factors contributing to reduced recoveries, including the need to train more staff and systemic problems in identifying certain kinds of discrepancies. For 2003–04, difficulty introducing a new computer system accounted for the lower recoveries. We estimate that in the two-year period, the matching program did not recover about $200 to $250 million in income taxes that it would have recovered without those difficulties. Agency officials advised us that, in 2005–06, the Agency plans to perform a second review of returns received in 2003–04 and verify a significant portion of returns not previously verified.

3.29 Excess RRSP contributions have not been detected. The accumulated amount of unused RRSP contributions that exceed a taxpayer's deduction limit by more than $2,000 is subject to a tax of one percent per month (under Part X.1 of the Income Tax Act). Reasons for excess contributions include

  • a lack of understanding of the rules,
  • failure to keep track of total contributions made in the year, and
  • a desire to shelter investment income for a number of years.

3.30 To date, the Agency has not taken steps to identify taxpayers who have made excess contributions but have not reported the amount of tax owed under Part X.1 of the Income Tax Act. Beginning with tax returns for 2004, the Agency plans to start comparing RRSP contributions reported by financial institutions with amounts reported by individuals on their tax returns. This will allow the Agency to verify whether taxpayers made the RRSP contributions for which they are claiming a deduction and to detect excess contributions. The Agency is studying the current extent of excess contributions to determine how to improve compliance in this area. These actions will address the concerns we expressed in our 1994 and 1997 reports about the lack of a verification program for RRSP contributions.

3.31 Recommendation. To improve the results of the matching program, the Agency should

  • fully document and support the basis for selecting returns to verify; and
  • verify returns in decreasing order of expected income tax recovery, in its work aimed at tax recovery.

Agency's response. The Agency agrees to fully document and support the basis for selecting returns to verify.

With respect to verifying returns in decreasing order of expected income tax recovery, the Agency agrees that case prioritization is important. However, it is not the only criterion that needs to be taken into consideration to produce better results for this business line.

The Agency will study how to improve its matching process to ensure that resources are expended on the best cases and is in the process of analyzing several alternatives to improve its current practices.

The Agency should make better use of certain information slips

3.32 Partnerships, issuers of flow-through shares, and promoters of tax shelters give the Agency information annually, detailing the amounts of income, losses, and credits attributed to their investors. The Agency uses information slips submitted by issuers of flow-through shares and promoters of tax shelters to evaluate the tax risk posed by particular issuers and promoters and to decide who should be audited. The Agency also uses this information during the audit to verify compliance with tax rules. The Agency uses information slips submitted by partnerships in a limited way, but not to evaluate tax at risk. However, the Agency does not use the information slips that all these independent sources issue when it looks at personal income tax returns to determine if claims for related losses and credits are valid or if all the related income is reported.

3.33 Tax legislation requires brokerage firms to report to the Agency on the proceeds paid or credited to clients from securities that they have disposed of or redeemed during the year. Though the Agency captures this information, it does not systematically verify whether taxpayers have included the proceeds from these transactions in the income or losses reported on their income tax returns. Auditors do use the information to screen files for possible audit, and the information is essential when the Agency audits a taxpayer. When tax officials analyze the information from brokerage firms, this can also result in special audits of taxpayers.

3.34 Recommendation. The Agency should develop and implement a strategy to make better use of information returns on tax shelters, partnership income, and flow-through shares in its compliance activities for personal income tax returns.

Agency's response. The Agency recognizes the importance of effectively utilizing information returns in its compliance activities. While the information received on tax shelter returns, flow-through share returns, and partnership returns is not designed to be automatically matched to taxpayers' files, the Agency does, as the Auditor General notes, use the information on tax shelter returns and on flow-through share returns to evaluate tax risk for audit purposes, and to verify compliance with tax laws. Concerning partnership returns, the Agency instituted a pilot project in April 2005 to assess the benefit of capturing partnership information into an audit database to improve risk ranking for audit purposes. The Agency will use the results of the pilot project to determine how best to utilize information returns in its compliance activities.

Measuring and reporting compliance

3.35 In its 2003–04 annual report to Parliament, the Agency states that it conducts random verifications in the processing review program. It goes on to say, "Based on the 2002 tax year, overall compliance for these items (key credits and deductions) remained high at 89.9 percent, a result which is comparable to last year."

3.36 However, when this information is compared with results over a longer period, as the Agency does in its internal program evaluation reports, and in its 2004–05 annual report, a different view of taxpayer compliance emerges. In verifying randomly selected returns, the processing review program found that the percentage of taxpayers claiming deductions and credits that they were not entitled to doubled from 5.42 to 11.08 percent between 1997 and 2003 (Exhibit 3.4). However, for the last four years, our estimate of the total revenue at risk has remained relatively constant. Through further analysis, program officials have determined that compliance rates appear to vary with the mode of filing.

3.37 The Agency needs to understand and address deteriorating taxpayer compliance in its processing review program. We were pleased to see that for deductions and credits covered by the processing review program, the Agency annually measures the percentage of taxpayers who did not comply with the rules. Where it has identified claims with a higher incidence of non-compliance, it has taken measures that include proposing changes to tax guides and other publications, updating guidance on the Agency's Internet site, and recommending changes in tax-filing software used by NETFILERS. In addition to this yearly analysis of program results, the Agency's compliance research program conducted a number of studies to learn more about the increasing rate of non-compliance in the general population and differences in the level of non-compliance among taxpayers, who use different filing methods, such as EFILE or NETFILE.

3.38 However, the Agency has yet to fully understand the reasons for the increase in non-compliance in items covered by the processing review program over the last seven years—whether the increase is attributable to specific deductions and credits, differences in taxpayer characteristics (for example, age, income range, province)—or is a general increase in non-compliance among taxpayers. While we recognize that the underlying causes of non-compliance in the processing review program cannot be easily "modelled" and explained, the Agency needs to answer these questions in order to address them with appropriate measures.

3.39 Measuring the dollar impact of non-compliance with rules for claiming deductions and credits. The Agency measures the percentage of taxpayers who are non-compliant in claiming deductions and credits that the processing review program verifies. Non-compliance can result from deliberate acts of tax evasion as well as from non-deliberate acts caused by honest mistakes in interpretation or mathematical errors.

3.40 However, the Agency does not measure the difference between actual federal, provincial, and territorial income taxes that are assessed and the estimated value of taxes that would be paid if 100 percent of taxpayers were to comply fully with the rules for claiming these deductions and credits. Using the Agency's data, we estimated, within $34 million, that federal, provincial, and territorial personal income tax revenue for 2002–03 would have been $586 million higher, if compliance had been 100 percent. By verifying deductions and credits in about three percent of the returns assessed that year, the Agency recovered $160 million, or 27 percent of the estimated shortfall.

3.41 We recognize that comparing actual tax revenue with an estimated value that would represent 100 percent compliance is a hypothetical exercise and that achieving perfect compliance is not possible. However, comparing actual results against the results for perfect compliance would allow the Agency to determine how well its processing review program identifies non-compliance and whether the dollar impact of non-compliance is worsening over time.

3.42 Measuring the impact of underreporting the income reported by third parties. As already noted, the matching program does not verify all returns in which discrepancies have been identified. The Agency does not analyze those returns to estimate the total amount of tax at risk for returns where discrepancies have been identified but not verified. However, if the Agency did this, it could compare the amount of tax at risk from unreported income subject to third-party reporting over a period of time.

3.43 Recommendation. To better understand trends in taxpayer compliance and to better measure and report on how effective the processing review and matching programs are, the Canada Revenue Agency should

  • report longer-term trend information on taxpayer compliance;
  • explain significant changes in compliance patterns and how it is addressing them;
  • report statistically valid information on the estimated tax impact of non-compliance with the rules for reporting deductions and credits, and for reporting income subject to third-party reporting; and
  • report on its performance in identifying and assessing the related amounts.

Agency's response. The Agency does currently report internally on long-term trend analysis, which includes identifying both changes in compliance patterns as well as action plans to address these changes. We will extend this reporting to ensure relevant information is included as part of the annual report to Parliament.

The Agency agrees to forecast potential revenue at risk for both the processing review and matching programs. We will commence estimating this value for these programs and also commence measuring our actual revenue recoveries against these estimates to broaden our program evaluation capacities. For the processing review program we are in a position to do so immediately (and in fact have done so effective for the 2004–05 Annual Report). However, the matching program recently implemented a new system that allows it to capture the required data. Therefore, as historical data is not currently available, it will be at least three to five years before we can estimate this value to an acceptable degree of accuracy.

Verifying income tax returns from domestic trusts

3.44 The Agency reviews all income tax returns from domestic trusts before it issues a notice of assessment. This review checks whether all amounts on the return are supported by the required documents and verifies the accuracy of calculations of the amounts reported on the documents. The Small and Medium Business Audit Program also audits domestic-trust income tax returns. Audits are performed in 45 tax services offices across Canada.

The Agency does not systematically evaluate the tax at risk in domestic trusts

3.45 The Income Tax Act provides for two broad categories of trusts  testamentary trusts and inter vivos trusts. In 2004–05 the Agency assessed 105,766 income tax returns of testamentary trusts and 71,150 returns of inter vivos trusts.

3.46 Tax officials audit a smaller portion of inter vivos trusts, as compared to testamentary trusts. Most trust audits in the last three years were of testamentary trusts selected from the trusts asking for clearance certificates. (Trustees often request a clearance certificate before winding up testamentary trusts and distributing property under their control to ensure that the trust retains enough property to pay any tax liability.)

3.47 Because the Agency lacks a systematic approach and sufficient information to evaluate the tax at risk in the entire population of domestic trusts, we looked to the tax recovered by audits as a rough measure of the risk of non-compliance. Over the last three years, the average tax recovered for each audit of an inter vivos trust was $22,300 compared with an average of $4,500 for each audit of a testamentary trust (Exhibit 3.5). These results suggest that the targeting of the Agency's audit effort may need to be re-examined.

3.48 Agency officials agree that they need more information about the amount of tax at risk in trusts. A pilot project was initiated recently to identify the nature and extent of non-compliance issues that are specific to domestic trusts.

3.49 The Agency is not provided with key information to evaluate level of risk. Financial information is one of the basic necessities for evaluating risk. Unlike corporations, trusts are not required to submit an annual statement of assets and liabilities. Information on the value and the kind of property held in a trust is important in evaluating the relative tax at risk for different kinds of property and for the trust as a whole. Given that trustees are required to report to beneficiaries on their stewardship of the property in a trust, it should not present an onerous compliance burden if the Agency requires a trustee to report a trust's assets and liabilities.

A number of factors have made trust review activities less effective

3.50 Information about assessors' corrections is not retained. Review procedures cover every box and line on the tax return but are generally limited to tracing amounts on the return to the required supporting documents and checking the calculations of the amounts on those documents. Currently, the Agency only measures performance in this area by whether tax officials process a return in the time allotted by the service standard. The Agency's target is to assess 95 percent of tax returns within four months of receiving them. Performance against this service standard is the only information in the Agency's annual performance report to Parliament about domestic trust tax returns.

3.51 Where necessary, an assessor reviewing a trust tax return will correct the information provided on the return. The corrections are recorded, and the taxpayer is notified of them through the notice of assessment. However, the system does not accumulate information on the corrections, including the tax changes, from these reviews. As a result, the Agency does not know the amount of additional tax revenue recovered by its verification activities. Nor does it have information showing where trusts have difficulty complying with the law—information that would help it better target its review efforts and develop appropriate activities to encourage compliance.

3.52 Income allocated to beneficiaries is not verified. Generally, a trust is taxed on income it earns unless the income is allocated to beneficiaries in the same year. When a trust reduces its taxable income by allocating income to beneficiaries, it must generally submit a T3 information slip to the Agency and to the beneficiaries showing the amounts allocated to each beneficiary. The beneficiaries must then include that income on their personal tax returns. This control is intended to ensure that either the trust or its beneficiaries are taxed on the income. However, the Agency does not currently verify that the deduction taken by the trust for allocation to beneficiaries matches the total income that the trust has reported on T3 information slips for those beneficiaries.

3.53 We asked tax officials to search the Agency's databases and determine the total income allocated to beneficiaries that trusts had deducted on their 2003 tax returns, and that officials believed would have resulted in a T3 information slip being issued. We also asked them to determine the total income allocated to beneficiaries on the T3 information slips they had received from trusts. We then compared the two totals. The numbers the Agency provided revealed a significant difference. When investigating this difference, Agency officials were able to correct amounts that accounted for a large part of the difference and offer plausible explanations for the remainder. While they are satisfied that a large part of the difference has no tax revenue implication, Agency officials need to complete their investigation and confirm both the reasons for the remaining differences and whether there are tax revenue implications. This finding shows that the Agency needs to put a compliance program in place to verify that the deductions claimed by a trust for allocation to beneficiaries are properly reported by the trust on information slips.

3.54 Requirement to submit complete information is not enforced. The Minister of National Revenue specifies which information trusts are required to report on the income tax return as well as any supplementary information. This information acts as an initial control to ensure amounts reported on trust income tax returns are accurate and complete, and it also enables the Agency to do a preliminary evaluation of whether trusts are complying with tax rules. We found instances where the Agency is not enforcing certain requirements to provide information. These practices make the controls less effective and undermine the requirement that the taxpayer file a complete and accurate return.

3.55 Established procedures are not followed. Tax rules require some trusts to be taxed on the increase in the value of their assets every 21 years. Assessors verify whether the rule applies to each return they review. If it applies, the assessor would expect to see the appropriate information included with the return so that the assessor can compare the attached information with the income reported on the tax return. To see whether assessors were following established procedures, we selected a sample of 20 trust tax returns for 2004 from a population of 451 that the Agency had identified where the 21-year rule might apply. We confirmed that the rule applied to eight of the twenty returns. In three of the eight cases the required information was not on file. Although this test was limited in size, such a high error rate shows that controls can be improved in this area.

3.56 Audit codes are of limited use. The review procedures that assessors complete are limited to a desk review of information provided by taxpayers. In certain circumstances the information provided could warrant that an auditor in the field do an in-depth review. Currently, the Agency asks assessors to input a number of codes in the system to note potential audit issues their review has identified. Our limited review of trust assessment files showed that assessors did not apply the codes consistently in cases where the assessment procedures required it. Auditors in tax services offices told us they could not determine from the automated trust system which trust returns have been given audit codes, nor did they receive referrals from the T3 Processing Section. As a result, the Agency is deciding which trust returns to audit without taking full advantage of information at its disposal.

3.57 Recommendation. The Agency should

  • systematically evaluate the tax at risk in domestic trusts when it selects income tax returns to verify;
  • require trusts to submit a statement of assets and liabilities with their trust income tax returns;
  • capture and compile information on corrections and develop reports to assist in, and account for, the management of trust assessing activities;
  • complete its examination of the amounts recorded in its databases for income allocated to beneficiaries on 2003 trust tax returns and on 2003 information slips and determine the tax revenue implications, if any, of the differences identified; and
  • implement a compliance program to ensure that domestic trusts are properly reporting the income they have allocated to beneficiaries on information slips.

Agency's response. Given the dynamic and evolving nature of trusts, the Agency is reviewing its approach to verifying domestic trust income tax returns, including analysis and consultations on the benefits and costs associated with requiring trusts to submit a statement of assets and liabilities with their income tax returns.

The Agency agrees with the recommendation to capture and compile information on corrections and develop reports to assist in, and account for, the management of trust assessing activities. In that regard, the Agency will build on its recently implemented Quality Evaluation Program to enhance its capacity to evaluate and manage trust assessing activities.

The Agency will complete its examination of the amounts recorded in the databases to ensure that there are no tax revenue implications.

Many situations do not require the filing of information slips. For all others, the Agency is evaluating this recommendation and the implications of more closely linking existing databases and enhancing the filing compliance of domestic trusts with respect to information slips.

Conclusion

3.58 The Agency's processing review program has a well-designed and well-executed risk-based approach for selecting and verifying deductions and credits that individuals have claimed on their tax returns but may not be fully entitled to. The program found that from 1997 to 2003 the percentage of such taxpayers doubled to 11 percent, although according to our estimate, the total revenue at risk has remained relatively constant for the last four years. Using the Agency's data, for 2002–03, we estimated that, had all taxpayers complied fully with the rules for claiming deductions and credits covered by the processing review program, federal, provincial, and territorial revenues from personal income tax that year would have been $586 million higher. Of that amount, the Agency recovered $160 million or 27 percent by verifying about three percent of returns assessed that year.

3.59 The Agency's matching program needs to improve how it manages risk when it verifies that taxpayers have reported the full income shown for them on information slips submitted by third parties. The Agency compares the information that taxpayers provided with information from third parties, identifies any discrepancies, and calculates the amount of income tax that is potentially recoverable for each return, where one or more discrepancies are found. While the matching program selects returns for review based on risk, its estimate of risk does not include the amounts the system calculated that are potentially recoverable. As a result, the Agency does not consistently select and review the returns with the largest amounts of potentially underreported income. Nor does it estimate the tax at risk for returns where it has identified discrepancies but not verified them.

3.60 In not monitoring the tax revenue impact of non-compliance in the areas verified by the processing review and matching programs, the Agency is not considering an important element in evaluating these programs' effectiveness.

3.61 The Agency does not systematically evaluate the tax revenue at risk in domestic trusts when choosing the tax returns it verifies. In addition, the lack of key information is hindering the development of an effective risk evaluation system; this key information includes such details as the value of assets and liabilities held in trusts. The Agency emphasizes audits of testamentary trusts. In the last three years, the average tax that the Agency has been able to recover from audits of testamentary trusts has been about five times smaller than average recoveries from audits of other trusts.

3.62 There are also deficiencies in the Agency's review activities for tax returns of domestic trusts. For example, the only measure of its performance in this area is whether a return was processed in the time that the Agency allots in its service standard; it lacks information on corrections that its assessors make; and it does not compare deductions claimed by trusts for allocation to beneficiaries against the amounts that the trust reported on the information slips provided to beneficiaries.

About the Audit

Objectives

The objective of the audit was to assess how well the Agency identifies and manages the risk that individuals and domestic trusts might fail to comply with the law when preparing their income tax returns, and how well it uses information from third parties in verifying personal income tax returns.

Scope, approach, and criteria

The audit covered verification programs for income tax returns of individuals and domestic trusts (other than registered trusts, such as RPPs and RRSPs).

For personal or individual income tax returns we focussed on the verification activities related to employment income, investment income, pension, and other passive income or losses and any other type of income reported on an information return and "line item" deductions and credits. We looked at how the Agency's seven verification programs interrelate to address all aspects of the tax return. We reviewed the extent, nature, and timing of the Agency's verification procedures for

  • income fields subject to matching with third-party slips;
  • selected claims, including the disability tax credit; and
  • the taxpayer's province of residence, marital status, and number of children.

We looked at the design and implementation of compliance measurement and file selection and at resourcing methodology for the confidence validity, processing review, and matching programs, and verified the accuracy of the information on audit results. We validated the Agency's information on compliance levels over the last seven years for deductions and credits subject to the processing review program, and we extrapolated the tax at risk in the entire population from the results of the Agency's verification of randomly selected returns for four years.

We reviewed the extent to which the Agency uses the following information slips submitted to it by third parties: T4, T4A, T3, T5, T5003 (tax shelter), T5008 (security transactions), T5013 (partnership income), and T101 (resource expenses). We also looked at the Agency's plans to match slips for RRSP contributions.

For domestic trust income tax returns, we looked at the pre-assessment review activities carried out at the Ottawa Technology Centre and the result of audit work performed by the Compliance Programs Branch for the previous three years. We discussed with Agency officials the scope and timing of a pilot project aimed at gathering sufficient information to develop a new audit strategy for domestic trusts.

Because we did not cover specialty audit programs such as tax avoidance and international tax, our audit did not cover the Agency's compliance activities for non-resident or foreign trusts.

We expected the Agency to

  • have adequate verification procedures for all the fields in a tax return,
  • rank tax returns for verification according to risk,
  • capture complete and accurate results of its verifications,
  • address known areas of non-compliance, and
  • make reasonable use of information slips that the law requires third parties to submit to the Agency.

These criteria were based on the Agency's annual performance report and related documents.

Audit team

Assistant Auditor General: John Rossetti
Principal: Barry Elkin
Directors: Denis Labelle (lead), John Pritchard

Wilson Ford
Catherine Johns
Sophia Khan

For information, please contact Communications at (613) 995-3708 or
1-888-761-5953 (toll-free).


Definitions:

Self-assessment—When a taxpayer completes an income tax return and determines the taxes owed based on his/her assessable income and allowable deductions and credits. (Back)

Non-compliance—When the taxpayer does not determine, report, or pay the correct amount of taxes owed. (Back)

Expected tax recovery—The amount that the system has calculated as the potential tax that can be recovered for a return multiplied by the probability that that amount will be recovered. (Back)

Testamentary trust—Is created, when an individual dies, to manage and ultimately distribute his or her property. (Back)

Inter vivos trust—Is created during a person's lifetime and includes all other trusts, such as mutual fund trusts and personal trusts. (Back)

Clearance certificate—A certification from the Canada Revenue Agency stating that all taxes and any related interest and penalties for which the trust is liable have been paid. (Back)