1994 Report of the Auditor General of Canada
Chapter 30—Revenue Canada—Goods and Services Tax: Audit and Special Investigations
Curtailment of "compliance test audits" and "first audits" results in lost opportunity to collect valuable information
Assistant Auditor General: Shahid Minto
Responsible Auditor: Jim Ralston
30.2 However, because some citizens fail to obey the law, well-targeted, credible sanctions against non-compliance are necessary. Non-compliance has significant effects: it results in loss of tax revenue needed to fund government programs and to service the national debt; it places those who comply at a competitive disadvantage; it contributes to an inequitable distribution of the burden of taxation; and it undermines the integrity of the tax system. When registrants do not remit GST collected from customers, they are misappropriating funds that belong to the government and creating a breach of trust.
30.3 From 1991-92 to 1993-94, there has been continued improvement in terms of additional tax revenue assessed as a result of GST audits. During this period, more than $2.1 billion was assessed by audit.
30.4 While significant, this amount was below projections. For example, in 1993-94, it was projected that audits, excluding those conducted in the province of Quebec, would generate $1.36 billion in additional tax assessments, but the actual amount was $613 million. The variance was due, in part, to projections that may have been too optimistic, fewer resources devoted to audits than planned, and lack of effectiveness in the audits that were carried out.
30.5 We reported in 1992 and 1993 that competing priorities have affected the development and implementation of enforcement activities such as registration and collection. In this chapter, we observe the impact on audit and investigations. Enforcement is impeded by a lack of information about registrants; by the absence of reliable tools for identifying and selecting registrants for audit; and by inadequate monitoring practices. Because of these and other factors, 40 percent of all audits conducted in 1993-94 did not result in additional tax assessments or reductions in refunds or rebates.
30.6 Raising additional tax assessments is one objective of audit. Another is to demonstrate enforcement efforts in the registrant population to remind people that they might be audited. During the last three years, the number of GST audits has declined, and the number of investigations and prosecutions has been low. Furthermore, close to one half of all audits consisted of limited checks conducted in GST offices, which do not generate significant awareness of enforcement among registrants.
30.8 Accordingly, a well-targeted, credible enforcement system is needed to detect and deter error, fraud and abuse. Enforcement activities are aimed at creating and encouraging compliance and demonstrating fair and consistent treatment of all Canadians.
30.9 In our previous reports, we pointed out the daunting task and difficult choices faced by Revenue Canada in implementing the GST. Ensuring compliance is a difficult task when consumers, businesses and government officials must become familiar with new legislation in an environment that seems increasingly hostile to compliance.
30.10 Revenue Canada is responsible for assessing, collecting or refunding GST for nearly two million registrants, many of them small businesses. At first, the Department's strategy to ensure compliance gave priority to education and assistance. This strategy recognized that the better informed people are, the greater the likelihood that they will comply with their obligations. But this approach is not without risk. Enforcement activities have suffered. As the GST matures, the emphasis of the strategy is shifting from education to other compliance activities.
Audit Objectives and Scope30.11 Our audit examined the adequacy of Revenue Canada's GST Audit and Special Investigations coverage, the results of its activities and the appropriateness of the support systems and procedures. We carried out our examination at headquarters and in seven district offices. To help the reader understand the context of our observations and recommendations, we have included in Exhibit 30.1 a description of audit and investigation functions typically applied in the administration of value-added taxes (VAT), of which the Canadian GST is one example.
Observations and Recommendations
Audit and Investigation Results
Additional tax assessments from audits are significant but less than projected30.12 Audits and investigations are designed to enforce compliance with the law and, by doing so, maintain public confidence in the integrity and fairness of the tax system. These enforcement activities must be conducted within a strategic framework that can adapt to newly identified threats or opportunities and form the basis for operating objectives and priorities. We therefore expect that the results will be assessed to see if objectives are met and priorities respected.
30.13 Additional tax assessments generated from audits have improved each year since the inception of the GST (see Exhibit 30.2 ). Revenue Canada has pursued two different audit strategies over the three-year period. In the fiscal years from 1991-92 to 1993-94, GST audits produced a total of more than $2.1 billion in additional tax assessments.
30.14 In 1991-92, the additional tax assessment from audits totalled $877 million and, as expected, was mostly attributable to reductions in Federal Sales Tax (FST) rebate claims. Having learned from the experience of New Zealand, Revenue Canada focussed on FST rebates and on GST returns where registrants were claiming refunds. New Zealand found that both rebates and refunds were subject to many fraudulent claims. FST rebates were meant to prevent double taxation by refunding FST paid on inventory held for resale. Revenue Canada wanted to ensure that legitimate, accurate FST rebates and GST refunds were processed and paid quickly and that accounts where there was suspicion of fraud or gross error were identified for audit.
30.15 In 1992, the Minister of Finance announced a series of measures directed to containing the deficit and revitalizing the economy. For Revenue Canada, this announcement translated into an enforcement strategy focussing on the generation of revenue, particularly from registrants who were not complying with filing requirements. In 1992-93, the total additional GST assessments resulting from audits amounted to $540 million, of which 25 percent came from non-filers.
30.16 The focus remained on revenue generation in 1993-94. The total additional tax assessments from audits came to approximately $723 million, including $110 million assessed in the province of Quebec by auditors from the ministère du Revenu du Québec on behalf of the federal government. However, the additional tax assessed was not as significant as projected: there was a variance of $700 million between actual and projected assessments from audits.
30.17 The variance can be attributed to three factors. First, it seems plausible that Revenue Canada projections may have been too optimistic, given that the Department had little or no historical data to go on. We have not tried to establish the reasonableness of the projections. Second, fewer resources were devoted to audits than had been planned. This is discussed in the following paragraph. Third, audits may not have been as effective as they could have been. Subsequent sections of this chapter point out ways in which audits could be improved.
30.18 The largest GST administration function in regional and district offices is audit, which, along with investigation, accounts for almost half of the human resources allocated to this tax. Since the GST started in 1991, many auditors and investigators have been diverted from their normal activities. During 1993-94, about 370 employees were diverted to other activities. Many switched to the collection function, and the Department estimates that the revenue generated in collections offset the revenue foregone in audit. The shortfall from projected audit tax assessments resulting from this diversion is estimated at $190 million. A similar workforce diversion in 1992-93 resulted in a shortfall of $133 million. No figure is available for 1991-92.
Audit coverage is low30.19 Although it is an important objective, the success of enforcement cannot be measured solely on the basis of additional tax assessments produced. Visible enforcement efforts to detect and correct non-compliance and to encourage future compliance are equally important. Audit is a vital component of the Department's enforcement activities and overall compliance measures. Given the current expression of public concern over evasion, we expected to find an increase in this enforcement activity. However during the last three years, the number of audits has declined. Moreover, as illustrated in Exhibit 30.3 , close to one half of all audits were desk audits. It is our belief that desk audits are less visible than field audits and generate less awareness of enforcement efforts among registrants.
30.20 Prior to the introduction of the GST, the Department developed a multi-year plan, anticipating that a mature audit program would be in place by fiscal year 1993-94. The plan was to have more frequent but shorter audits than under the FST regime and to rely on deterrence to a much greater extent. Key elements of the audit strategy were the volume of audits to be conducted and the number of audit staff required. It is important for the Department to be able to perform its audits promptly, since an assessment or reassessment of a registrant must generally be issued within four years after the day on which the taxes became payable. Assuming a population of 1,600,000 registrants, it was expected that 152,000 audits (excluding audits of FST rebates) would be performed every year by about 2,300 auditors.
30.21 Exhibit 30.4 presents the number of audits completed since the GST was introduced. The exhibit shows that after the first year, the Department failed to meet its target of 152,000 audits. In 1993-94, about 80,300 audits were completed. It should be noted that, as discussed in a subsequent section, even this number of audits is overstated due to discrepancies within the GST automated system.
30.22 Comparison with other countries provides another perspective from which to assess audit coverage in Canada. In 1992-93, the United Kingdom had 1,600,000 registrants and performed 378,000 audit visits. Granted that the type of intervention (see Exhibit 30.1) and number of auditors may differ from one country to another, this level of activity is more than four times the Canadian level.
30.23 In our 1989 Report, we expressed concern that Revenue Canada had shown record low rates of Federal Sales Tax audit coverage - less than seven percent at that time - and that there was a continuing downward trend in the ratio. In a report on income tax enforcement programs in 1990, we expressed a similar concern with the decline in income tax audits. We have the same concern this year with respect to GST audits, for which the current coverage rate is declining toward three percent.
Number of investigations and prosecutions is low30.24 The results of GST investigation activity started to become apparent in 1993-94. During that year, 107 cases were closed after preliminary investigations, and 17 cases were closed after detailed investigations. There were 15 convictions, five of which resulted in jail terms and hefty fines. As of March 1994, the Department had 240 cases in the investigation workload, and there were 33 cases before the courts.
30.25 The number of detailed investigations and prosecutions was less than 50 percent of the planned figure. The reasons for this low level of investigation activity are threefold. First, it is very difficult to start with a strong investigation unit at the beginning of a tax regime, since most cases are obtained through complaints and referrals. It takes some time to build the necessary information and to train investigators. It may also take one to two years to complete a case. Second, while the audit function struggled to establish its own foundation during the past three years, expectations in terms of quality and number of referrals to investigations were not clear. Furthermore, the type of audit intervention performed, such as desk audits and field audits with a high focus on quick generation of revenue, reduced the potential for "good" referrals to investigations. Third, fewer resources were used in investigation activities than planned. For example in 1993-94, there were 171 employees planned for GST investigations; however, only 149 were used in this function.
Federal Sales Tax audit coverage was lower than expected30.26 In 1989, we indicated that a strong close-out strategy and efficient deployment of audit resources would be necessary during the transition from the Federal Sales Tax (FST) to the GST. An internal review was initiated by the Department in 1992 to assess the final phase of the FST licensee audit program. Preliminary findings indicated that audit coverage was not as high as expected during the transition period, since resources were shifted to implement the GST.
Audit and Investigation Systems and Practices30.27 Enforcement activities should be conducted efficiently, fairly and consistently. In an environment of scarce resources, where risk of fraud and error is high, sufficient and appropriate management information is crucial to the success of an enforcement program. Equally important are adequate planning, supervision and monitoring.
30.28 We found that the lack of information on which to base the selection of registrants for audit is a major impediment to district offices' ability to target enforcement actions effectively. We noted in our 1992 Report that the Department decided to limit the application for registration to a one-page form to simplify its completion, but at the same time, this reduced enforcement capability. The limitations described below concerning the recording and analysis of audit information have also restricted the efficiency and effectiveness of the audit function.
System discrepancies compromise accuracy and reliability of reported results30.29 Information to manage the audit program is derived from the GST automated system. Information is captured at registration time and when registrants submit their returns and payments. Upon completion of an audit, the GST automated system is updated with selected information, including the period audited, amounts assessed, time required to complete the audit and revised registrant information. From this system, summary reports are printed monthly, showing the type, number, total audit time and revenue change for completed audits. While these summary reports are useful, they do not permit analysis of audit results by categories such as industry sectors, business size, registrant location or "no-change" audits. Therefore, ad hoc reports were developed to supplement these summary reports. However, the production of these reports has often required extensive lead time and carried risk of errors and inconsistencies.
30.30 In September 1993, headquarters and the district offices received a copy of the audit information for 1993-94 contained in the GST automated system to analyze on stand-alone computers. This practice reduced the need for ad hoc reports. We obtained and analyzed these same data and noted discrepancies that compromise the accuracy and reliability of the reported audit results. For example, the GST automated system allows the same audit to be recorded more than once and counts corrections of the Department's own record-keeping errors as completed audits. As a result, the number of audits reported by the GST automated system is overstated.
Information to target audit selection is incomplete30.31 All types of audit require tools or techniques that will ensure effective and consistent selection across the country. Because certain registrant information is not collected at time of registration, initiatives to develop needed information for such tools have had only modest success.
30.32 We reported in 1992 that the registration form lacked certain data needed for administration of the GST and that these data should be obtained when registrants are first contacted. A primary example is the Standard Industrial Classification (SIC) code that is assigned to registrants, describing their business sectors and main activities. The code is essential for targeting audit selection, studying compliance behaviour and measuring revenue generation in specific sectors. In 1992, about 700,000 accounts were not coded. At the end of March 1994, there were 420,000 accounts still not coded, varying from 3 percent of the registrant population in one region to 36 percent in another. Starting in February 1994, measures have been taken to ensure that a valid SIC code is entered when updating the GST automated system for audit results. However, up to that time, we observed that in 10,000 cases the SIC code information was still missing after an audit had been conducted.
30.33 The development of a system that assigns points to registrants based on criteria such as annual sales, industry and time since last audit began in the fall of 1992. A list of registrants with high scores was generated and delivered to the district offices in March 1993. The development of this point-scoring system was adversely affected by a number of factors. The analysis was based on a small sample of primarily GST refund claims, which included a high number of desk audits. The points assigned in this system were developed on the basis of data that were incomplete due to missing returns, missing annual sales figures and missing SIC codes. The Department continued attempts to improve this selection tool, and another list of high-scoring registrants was sent to the district offices in May 1994. At the time of our audit, the extent of use and results from this new list were not known.
30.34 District offices also have access to other information sources for audit selection: selection systems developed in regions (similar to the national point-scoring system); the registrant data base created from the GST automated system; lists of non-filers; referrals from registration and special investigation units; and examination of business advertisements, title searches and business listings in the yellow pages of the telephone directories. Capturing this information on the GST automated system would permit analysis of the comparative effectiveness of these sources or methods of audit selection.
Curtailment of "compliance test audits" and "first audits" results in lost opportunity to collect valuable information30.35 Compliance test audits are an important mechanism for obtaining information to manage the audit function. These tests involve auditing a random sample of registrants in various industry sectors. They are full-scope field audits that can help draw conclusions about the level of compliance and the revenue gap. In 1992-93, the intent was that 35 percent of audit effort would be directed toward compliance test audits; the results would be used to provide an early warning system for problem sectors and to direct further audit activity to areas of high risk and areas where audits could be expected to deter non-compliance.
30.36 Compliance test audits started as planned in 1992-93 but were abandoned six months later. The Department stated that this move was made in response to the government's concern relating to the collection of overdue taxes. Compliance test audits generated, on average, a lower immediate impact on tax assessment than did other types of audits; the switch in strategy was definitely beneficial from this point of view. Less clear, however, is the long-term effect of not collecting essential information required to improve automated verification procedures, computer-assisted audit selection techniques and the capacity for early detection of changes in compliance behaviour.
30.37 The term "first audit" refers to a visit to a new registrant in order to explain obligations, to detect businesses registered for fraudulent purposes and, generally, to detect problems early, when they can be easily corrected. Revenue Canada conducts virtually no first audits, although it originally planned to do approximately 14,000. The Department stated that first audits were abandoned because it felt that, in a period of limited resources, the resources could be more effectively used in other audit programs. Consequently, an excellent opportunity was missed to gather basic information useful for future audit, registration or collection actions-information such as details of taxable and exempt activities, annual sales, quality of accounting records and systems, related corporations, validity of elections filed and bank accounts.
Screening returns for audit takes more time than planned30.38 To understand how returns are selected for audit, one must begin with some facts on how returns are processed. GST returns and rebates are processed at the Prince Edward Island and Quebec processing centres. In 1993-94, about 7.3 million GST returns and 300,000 GST rebates were received. Of the GST returns, 2.3 million claimed refunds (credit returns) and 5 million owed money (debit returns). Exhibit 30.5 shows the volume and flow of processing and auditing GST returns and rebates for 1993-94.
30.39 A prepayment audit system is in place to prevent payments of rebates and refunds to registrants where there is suspicion of fraud or gross error. The system consists of a series of automatic validity checks, including dollar thresholds for individual amounts claimed, cumulative thresholds for payments to date, separate checks for registrants who have never been audited or who are submitting their first claim, and checks for inconsistent information on returns or applications. Returns and rebates that pass the validity checks are automatically paid. In 1993-94, approximately 630,000 claims failed validity checks and entered the audit workload. We noted that little analysis had been done with respect to the usefulness of the various validity checks. Auditors are to establish thresholds tailored to individual registrants in order to prevent recurring credits from entering the audit workload. We found that registrants' individual limits had not always been placed on registrant accounts, despite the national effort to do so. In fact, even after a refund or rebate audit, individual thresholds were not recorded by the auditor in 7,016 registrant accounts in 1993-94.
30.40 Returns that fail the validity check are reviewed through a desk review process, and those perceived as presenting a risk of gross error or fraud are selected for audit. The desk review process is very labour-intensive, involving much of the work time of audit supervisors. The Department accounts for time devoted to desk reviews under the caption of supervision, training and administration. For 1993-94, the target for this category of time usage was set at 34 percent of auditors' time, but, in fact, 47 percent of auditors' time was taken up by such activities.
Validation of complaints and referrals could be done more efficiently and effectively30.41 During 1993-94, Special Investigations received and took action on 4,000 complaints and referrals from the public and from various internal sources in connection with suspected violations of the Excise Tax Act . Because of the risk and volume involved, it was important that they be dealt with efficiently and effectively.
30.42 Forty-six employees were used to validate the information and to determine whether enforcement action should be taken. Approximately 1,800 of the 4,000 cases were referred to other functional units such as Audit or Registration for further examination. The remaining 2,200 cases were discarded from further enforcement action because they lacked substance.
30.43 Having investigators perform this function reduces the time they have available to devote to actual investigations. Possible alternatives include:
- providing Audit or other functional units within the Department with training or guidelines on how to improve the quality of referrals from these units;
- freeing investigators from preliminary review tasks by allocating as much of this work as is practical to staff at lower levels.
Audit and investigation work is not being adequately monitored30.44 Appropriate supervision and monitoring mechanisms are necessary to ensure that audit and investigation decisions are consistent. These mechanisms include file reviews after an audit, on-the-job training, management information systems and quality review of district audits and investigations. We noted that there is a need for closer supervision and monitoring of the work performed by GST auditors in each district office.
30.45 Auditors face, with increasing frequency, complex issues raised by tax specialists who are often interested in testing the application of the legislation in certain circumstances for the benefit of their company or their clients. The changes in priorities, amendments to the legislation, diversion to other functions and extensive time spent in workload management activities have made it more difficult for GST auditors to gain the experience needed to deal with complex issues. In view of this situation, audit supervisors need to be more involved in the planning and review of auditors' work and decisions and need to promote the exchange of lessons learned about the behaviour of certain sectors of the registrant population and the avoidance schemes that are in use.
30.46 Completed audits that we looked at did not show evidence of pre-audit planning or post-audit review. The scope, time and tools used to conduct the same types of audits varied considerably. We also noted that quality assurance and monitoring by regional offices of audit and investigation work in district offices have been sporadic. These circumstances have contributed to variances among regions in the number of desk audits as a percentage of total audits (see Exhibit 30.3 ) and in the number of audits with no change in tax assessed (see Exhibit 30.6 ). Approximately 31,000 audits resulted in no change in tax assessed in 1993-94.
30.47 We noted that procedures followed and information captured in certain cases were not adequate and that training provided to special investigators needs to be strengthened. Investigators have not yet had extensive experience in conducting investigations; therefore, their activities and decisions need to be closely monitored. A case could be lost if the investigator does not provide sufficient evidence, although so far the Department's record in obtaining prosecutions is very good.
30.48 There is a need for stricter control, co-ordination and monitoring of audit and investigation activities, both to improve the current results and to guarantee the application of minimum standards to ensure fair treatment of registrants and consistency in enforcement decisions.
Audit Laptop System shows promising results30.49 All GST auditors have been provided with laptop computers and have been trained to use the Audit Laptop System (ALS). ALS is a set of audit tools. It provides an extensive number of audit programs and procedures, such as the comprehensive audit program, the GST refunds audit program, procedures for rebates and procedures for audits of financial institutions. It includes audit manuals, some industry sector profiles, departmental directives and policy papers, GST legislation and memoranda, bulletins, word-processing and spreadsheet facilities.
30.50 Providing auditors with access to many sources of information helps reduce the time needed for research. In addition, the provision of predetermined audit programs and procedures assists the auditors in adhering to work standards and in carrying out their responsibilities in an orderly fashion.
Accountability and Reporting
Key performance indicators are not reported to Parliament30.51 Audit and Special Investigations have specific roles to play in the overall compliance and enforcement strategy. The audit function is meant to verify compliance and maintain public confidence in the integrity of the tax system; the investigation function is aimed at the detection and deterrence of fraudulent activity. Given the significance of their roles, we would expect relevant, reliable and understandable information on these functions to be reported to Parliament.
30.52 As discussed earlier, comprehensive analysis of audits and investigations has been limited because of lack of information and system discrepancies. Accordingly, performance information reported in Part III of the Estimates has been limited too, presenting information on workload only. In our view, as it currently stands, the information provided on these activities could be enhanced by including:
- a breakdown of the number of activities by type of intervention. There are significant differences in terms of impact and resources required for desk audits versus field audits, for audits of small businesses versus those of large corporations or for preliminary investigations versus detailed investigations.
- the utilization of human resources. Since close to 50 percent of the GST staff are employed in Audit and Special Investigations, indicators of efficiency (such as performance in relation to the time standards per audit or the use of audit resources in indirect activities) would be relevant to report.
- the coverage rate of the population and the impact of audits and investigations. Information regarding the efforts and success of the Department in identifying and targeting high-risk areas and the impact of these interventions could also be reported. For example, in order to help assess the impact of audit interventions, it would be useful to disclose information on additional tax assessed, interest and penalties charged, fines and jail sentences imposed and the reduction in amounts claimed. Comparisons with targets and previous years' achievements would also be useful.
Conclusion and Recommendations30.53 This chapter repeats a theme that appeared in our previous work on the GST: since inception of this tax, the capacity of the Department to ensure compliance has been seriously challenged. In 1992, after the introduction of the GST, we noted that the Department's capacity to identify new registrants was limited. Furthermore, we observed that enforcement activities had been curtailed due to redeployment of resources to other activities, such as providing information to registrants. While examining the status of GST collections in 1993, we discovered a high level of non-compliance: almost one in three registrants was delinquent in some fashion. We recommended that all essential aspects of the departmental collection strategy be reviewed, including the allocation of resources to this particular enforcement activity.
30.54 Because the GST was new, auditors and investigators had to go through a learning curve and adjust their practices and procedures in response to changes in legislation and in priorities. This situation was expected, but it reduced the extent and potential impact of audit and investigation activities.
30.55 Revenue Canada must deal with a registrant population larger than expected and with significant non-compliance risks. To do so, it must achieve a greater level of maturity and stability in enforcement activities. The recommendations that follow are addressed to Revenue Canada and are aimed at improving the coverage and the results of audit and investigation activities, and the practices and procedures that support these activities. These recommendations are also aimed at ensuring that, in the future, members of Parliament will be provided with relevant, reliable and understandable information to enable them to assess enforcement efforts and measure the threat of loss of tax revenue needed to finance government programs and to service the national debt.
30.56 Revenue Canada should pursue its initiatives outlined in Exhibit 30.7 in order to:
- improve the quality and quantity of registrant information it possesses;
- determine objectives and implement operational plans to ensure visible enforcement efforts throughout the registrant population and to obtain greater amounts of additional tax; and
- develop effective audit selection techniques, in part by applying the results of experience acquired in various regional and district offices.
- review the efficiency and effectiveness of its present methods of validating complaints and referrals to Special Investigations; and,
- improve its reporting in Part III of the Estimates by including, at a minimum, performance information on the number of activities by level of intervention, utilization of human resources, coverage rate and impact of audits and investigations.
The Department supports the need for increased audit visibility to effectively ensure good registrant compliance. As Exhibit 30.7 notes, the administrative consolidation of the Department is seen as key, in achieving improved cost-effective compliance. Harmonized audit activity is already proving effective and its effectiveness in the context of visibility and compliance will increase with an integrated approach to audits and investigations under the Customs Act, the Excise Tax Act, and the Income Tax Act.
The Department is continually refining its audit selection techniques as indicated in Exhibit 30.7 and will be taking additional steps to harness its experience in assessing the effectiveness of the audit selection system.
The Department initiated, in May 1994, a review to determine the best process to validate complaints and referrals to GST Special Investigations and also to maximize resource effectiveness. The Department is taking steps to implement the results of the review.
The Department will undertake a review of reporting in Part III of the Estimates to include the various elements noted in the recommendation.