4.1 As outlined in the previous chapter, benchmarks were generally considered by stakeholders to be a useful ATO risk identification tool for non-compliance in that they excluded a large number of likely compliant taxpayers from ATO compliance activities.

4.2 However, stakeholders also raised strong concerns about the use of it as a risk identification tool in isolation because they considered that many compliant taxpayers were selected for compliance verification activities which impose significant costs.

4.3 They were of the view that being significantly outside the benchmarks was, in and of itself, not enough reason to warrant compliance verification, and that additional indicators should be considered to determine whether the risk of underreported income was sufficient to commence action which imposed a significant compliance burden.

4.4 In particular, stakeholders considered that, as far as reasonably possible, the ATO's correspondence audits should be improved to provide for the use of reliable predictors of non-compliance before the imposition of a costly compliance burden such as being required to send numerous records to the ATO.

4.5 These issues outlined above are discussed in more detail below.

Variance from the benchmark as an indicator of underreported income

4.6 Under the benchmarking strategy, where a taxpayer's variance from the benchmark reaches a certain threshold and certain exclusions do not apply, they will be automatically selected for compliance activities. That is, the degree of variance from the benchmarks is the primary means of case selection, including phone reviews and correspondence audits.101 However, it should be noted that the ATO has not undertaken any analysis to establish whether such variance is indicative of the level of risk to the revenue. For example, it does not use its existing correspondence audit results to compare the business's variance from the benchmark with audit outcome results.102

4.7 In relation to phone reviews, stakeholders have provided positive feedback on the ATO's most current approach. These reviews seem to be quick and involve minimal preparation — two to three hours as reported by some tax practitioners.

4.8 However, those stakeholders involved in correspondence audits provided significant feedback on the conduct of those audits and, in particular, what they observed as compliant taxpayers being wrongly selected. Submissions indicated that a large number of cases were, after varying degrees of intensity, concluded with no further action. For some representative groups, the targeting of correspondence audits was perceived to be largely 'random' or a 'scattergun' approach.

Compliance costs

4.9 Correspondence audits usually require the taxpayer or their tax agent to provide hard copies of records, usually for a nominated quarter of the year in review (often April-June). This is one of the more intense high volume ATO compliance activities conducted under the ATO's benchmarking strategy.

4.10 The general contention in submissions to this review was that these audits impose significant compliance costs on taxpayers and tax agents. It is important to appreciate that a baseline level of compliance costs is expected to be borne by taxpayers as a feature of the self-assessment system. The particular concern raised was in relation to increased compliance costs that could have otherwise been avoided or minimised.

4.11 Where a taxpayer is meticulous with its record keeping it incurs 'baseline' compliance costs — that is those costs involved in recording and maintaining evidence of income and expenses, and providing to the ATO access to these records. In relation to a desk audit additional costs are incurred in collating and transmitting these records to the ATO — such as printing out the records and freighting the boxes to the relevant ATO office (where the ATO does not receive electronic transmissions of records). Taxpayers and tax agents also reported difficulty and added expense in gathering and supplying certain records in the timeframes provided for a range of reasons.

4.12 A few examples are outlined below that reflect some of the more costly cases brought to the IGT's attention.

4.13 One tax agent wrote responses on behalf of all his clients who received a letter from the initial bulk mail out letters campaign mentioned earlier. He did not charge for these letters, which took on average about 45 minutes each.103 Another tax agent charged $8000 to handle an audit due to the amount of work involved, while another agent charged $4500 to handle another audit ($180 per hour for more than 20 hours' work).

4.14 As a result of these costs, some tax agents turn away clients who the ATO consider may have record keeping concerns, since the agent believes they will not be able to recover the fees from them.

4.15 One representative group said:

Some practitioners find these services difficult to recoup fees from, as client resist paying for what they consider unnecessary compliance activities. This is especially in cases where the benchmark is poorly targeted in the first place.

4.16 Tax agents generally felt that the ATO has effectively 'outsourced' a lot of the auditing administration work to the tax agent. This is especially the case in terms of collating and sending records to the ATO. A range of stakeholders felt it would have been better and cheaper for them to have a field audit instead.

4.17 It is apparent that the conduct of the auditor has an impact on the level of compliance costs borne by the taxpayer. This varies with respect to the amount of research undertaken during the beginning of an audit.

4.18 One case reviewed by the IGT is illustrative of conduct that reduces taxpayers' compliance costs. In a correspondence audit, the auditor called the tax agent as part of their research. They established that whilst they were 'roughly' in the same industry of 'roofing services', the actual business did not reflect the description of the industry in the benchmark. That is, the benchmark applies to businesses that mainly install tiles, and does some restoration and plumbing. The taxpayer in question only did restoration. As a consequence of the auditor's discussion with the taxpayer, the case was withdrawn from the program before the standard audit letter was sent.

4.19 In another case, by way of contrast, a dam maintenance technician was incorrectly classified as a roofing services provider. The initial auditor clarified that roofing was the wrong benchmark to use, but then sought to apply the plumbing benchmark in substitution. An audit then proceeded on the basis that the taxpayer was now outside the plumbing benchmarks. After 5 months of communication between the taxpayer and auditor about getting records together, the case was then reallocated to another auditor, who determined that the plumbing benchmark was not appropriate for dam maintenance. This auditor established that there was in fact no benchmark for this type of taxpayer and closed the case with no further action on that basis within days of receiving the case.

4.20 Stakeholders argued that due to the costs imposed, it was important that the correspondence audits were targeted towards likely non-compliant taxpayers only.

Strike rates

4.21 Strike rates are one means to examine whether the current case selection methodology is effective at targeting non-compliant taxpayers. They can also be useful in understanding whether compliant taxpayers have been identified incorrectly. A 'strike rate', as the name implies, is the proportion of completed cases that resulted in an increase in income tax or GST divided by the total number of cases selected and conducted (that is the total number of cases, excluding early exit cases). As previously noted, these are also known as outcome cases, where there is a 'strike'.

4.22 An overview of ATO strike rates is provided in the table below and is compiled by the IGT from ATO sources. Strike rates for some of the ATO compliance products are discussed thereafter.

Table 13: Strike rate for benchmarking activities from May 2010 to April 2011
Intensity Product Outcome cases Total completed cases Planned strike rate(a) Actual strike rate Liabilities raised(b)
Highest (field) Cash Economy Audit 143 276 60% 52% $5,856,166
(field) Specific Audit 102 143 70% 71% $2,552,445
(desk) Correspondence Audit 1879 7670 32% 24% $57,097,719
(field) Record keeping audit 104 927 40% 11% $652,301
Lowest (desk) Phone review 10 9839 0% <1% $117,358

Source: Figures compiled by IGT from ATO spread sheet data supplied 23 December 2011 and 31 May 2012.

Note (a): Planned strike rates taken from 2011-12 Cash Economy Staff Allocation and Case Numbers supplied by ATO 23 February 2012.

Note (b): Liabilities include GST, Income Tax, PAYG withholding and PAYG instalments, but excludes penalties.

Phone reviews

4.23 As set out in the table above, from May 2010 until April 2012, the ATO conducted 9839 phone reviews. Of these, 9829 (99 per cent) resulted in no further action whilst less than one per cent resulted in an outcome. Since phone reviews are primarily designed to gather information and to increase the taxpayer's awareness of record keeping obligations rather than to check if income is correctly reported no strike is expected.

Record keeping audits

4.24 During the same period, the ATO conducted 927 record keeping audits. Of these, 104 cases resulted in the application of record keeping penalties or adjustments to income or GST liabilities.

Correspondence audits

4.25 From May 2010 until April 2012, the ATO conducted 7670 correspondence audits. Of these, 24 per cent resulted in an outcome, being an amended assessment on some basis agreed to by the taxpayer or a default assessment. The other 76 per cent resulted in no further action.

4.26 The IGT has inspected a sample of nil outcome correspondence audit cases. The most common reasons for recording these cases as a 'nil outcome' were:

  • After a phone conversation between the ATO officer and the taxpayer or their agent, the ATO officer determined that the taxpayer was in the wrong ATO benchmarked industry and was either within the benchmark range of the correct ATO benchmarked industry or no benchmark existed for the industry to which the business belonged.
  • After the commencement of the audit, the ATO officer and taxpayer discussed the fact that the taxpayer had a mixed business and selected the main business for their 5-digit ATO business industry code. The taxpayer supplied some documentation disaggregating their business into their separate lines. The ATO officer, after ascertaining the ratios for each business line, determined they were all within the benchmarks. Alternatively, one business line was out of the benchmarks, but the amount at risk was low. Depending on whether the business already keeps disaggregated business records, this type of case can have a lower or higher compliance cost. Smaller businesses do not tend to keep disaggregated accounts and would therefore need to expend additional time or money to disaggregate them for the ATO.
  • The taxpayer was in the correct industry, they supplied all required records, and the auditor was satisfied that the records evidenced the figures reported in their BASs and income tax return. That is, the audit ran its full course and was found to be compliant.

Specific audits

4.27 As indicated in chapter 2, most benchmarking specific audits flow on from other compliance activities. For example, where an ATO correspondence audit on a partnership has resulted in an adjustment to the sales figures of the partnership business, it will necessarily require an adjustment to the income distributions of the partners. The specific audit is the ATO compliance product used to determine that adjustment. As such, higher strike rates are expected. They are not selected on the basis of variance from the benchmark.

Cash economy audits

4.28 The highest intensity benchmarking audit product is the 'Cash Economy Audit'. Under the benchmarking strategy, this involves ATO officers visiting the taxpayer's business premises (a field presence). The ATO conducted 188 cash economy audits over this same period, with 55 per cent resulting in an outcome or strike rate from an ATO perspective.

4.29 Under the benchmarking strategy, these cases may be selected as a result of either a significant variance from the benchmark in addition to being identified as a risk under the cash economy risk model or escalated by an ATO officer from a correspondence audit.

IGT observations

4.30 The ATO's case selection methodology appears to hypothesise that a higher variance from the benchmark presents a higher underreporting risk than those with a lower variance. Cases which have a variance below a certain threshold may be selected for a record keeping audit, whilst other cases above the threshold are excluded from record keeping audits, but may be instead selected for the more intense correspondence audit.

4.31 It should be noted that the process the ATO undertook to identify cases selected for correspondence audit involved a significant amount of filtering, starting from 1.4 million potential cash economy taxpayers. From this figure, the ATO has used the benchmarks to select 7670 taxpayers for correspondence audits from May 2010 to April 2012.

4.32 As correspondence audits have a strike rate of 24 per cent this may suggest that a higher variance from the benchmark does not necessarily, of itself, indicate likely underreporting of income. This is not to say that the variance from the benchmark is ineffective at targeting a proportion of underreported income. As a starting point for risk hypothesis testing, benchmarks have identified non-compliant taxpayers.

4.33 However, it is equally true that correspondence audits may be better targeted by not commencing them solely based on benchmarks. The correspondence audit strike rate of 24 per cent is better than auditing a random sample of the 1.4 million cash economy population. However, the IGT believes that the current approach may be improved in better targeting its correspondence audits towards likely non-compliant taxpayers.

4.34 The IGT notes that the overall strike rate may be affected by a lack of distinction between nil outcome cases where audits are closed early as opposed to where the audit runs its course and the taxpayer is found to be compliant. Classifying both of these types of cases as 'nil outcome' prevents the ATO from understanding the true extent to which audit nil outcomes can be attributed to either incorrect case selection or taxpayer compliance with the law.

4.35 The ATO should consider reporting their cases in such a way as to make a distinction between those cases which are closed due to incorrect case selection shortly after taxpayer contact is made and those cases where the risk hypothesis is tested by the ATO.

4.36 It is the IGT's view that that the ATO could do more to bolster the risk identification process before commencing a correspondence audit, with its associated compliance costs and intensity. There may be many reasons why a business falls outside of the benchmarks that are not related to non-compliance with the tax laws. For example, some taxpayers may be simply better at their business than others and there may be different personal motivations affecting costs (such as wanting to use better quality materials, employing more staff because they want to spend more time with their family, etc.).

4.37 However, regardless of the taxpayer's motivations for running their business, they are still required to maintain adequate evidence of their reported income and expenses. The challenge for the ATO is to use a method to identify potential underreported income, which may ultimately be evidenced by inadequate record keeping. Benchmarks are but one indicator and the ATO should use other indicators to refine the targeting of its compliance verification activities away from compliant taxpayers and towards more likely non-compliance taxpayers.

Other indicators of underreported income

4.38 Stakeholders expressed a variety of views about what these other indicators were, and the means by which the ATO could access information relevant to those indicators.

4.39 Firstly, stakeholders suggested that the presence or lack of cash controls may point to underreported income, and that the ATO could infer the presence of such controls by whether the business employs staff with access to cash. If so, the business owner has a stronger incentive to have cash controls in place, to minimise theft. Conversely, if the business is family run such cash controls may be less likely to be in place. It was considered that the ATO could identify such businesses from employment declarations lodged with the ATO without the need to contact taxpayers.

4.40 Secondly, it was suggested that businesses predominantly dealing with other businesses (that is 'business-to-business') are more likely to have better record keeping since their business customers would normally require invoices.104 Furthermore, it was also considered that business customers are more likely to pay electronically compared to retail customers who are more likely to pay in cash. However, it was acknowledged that the ATO could have difficulties in identifying such taxpayers without verifying their records first.

4.41 Thirdly, stakeholders expressed the view that the 'tax competency' of the business operator may have a bearing on their compliance105 and, as such, the ATO should consider this as an additional step in the risk identification process. Such tax competency could be affected by the business owner employing a bookkeeper to maintain its records.

4.42 Stakeholders also suggested that another indicator of underreported income is where the reported business income is unrealistic given the level of taxpayer assets, household expenses or living standards. In this respect, where such business income appears unrealistically low, the ATO could make enquiries about whether the taxpayer has other non-business income or if someone else is financially contributing to the household. To a certain extent, the ATO does this through its cash economy risk model.

4.43 There was a variety of stakeholder views, however, on how the ATO may best obtain information on these factors. For example, the business may need to complete additional fields in a tax return on a regular basis or the ATO may contact the business by phone to make additional enquiries. In this respect, it was acknowledged that there may be difficulties in either redesigning income tax returns for a small proportion of the total taxpaying population or evidencing taxpayers' representations over the phone without imposing substantial additional compliance costs.

4.44 The ATO itself is also aware of a range of indicators which may potentially point to a higher risk of underreported income in the cash economy, including:

  • lack of proper record keeping practices; and
  • poor lodgment history.106

4.45 Further, the ATO is aware that large franchise retailers are a low record keeping risk since the franchisors have computer record keeping systems in place.107

4.46 The ATO also advises that it excludes classes of taxpayers from compliance activities on certain factors. None of these factors, however, relate to any of the indicators described above.108

4.47 The following table shows how the various indicators mentioned above are used or not used by the ATO when deciding to commence a correspondence audit. In the third column, the ATO has advised the indicators for which it has ready access.

Table 14: Indicators of underreported income used or not used by the ATO for correspondence audits

Indicators referred to by stakeholders and the ATO

Indicators currently used by ATO in deciding to commence correspondence audit

Indicators the ATO can use

The presence and adequacy of cash controls No No
Whether the business employs staff No Yes
Whether the business has access to cash Yes. Note the ATO selects businesses in cash economy industries. Yes
Whether the business mainly deals with consumers or other businesses No No
The skill of the person responsible for record keeping No No
Whether the business engages the services of a tax agent and/or bookkeeper No Yes
Whether the business is significantly outside the benchmarks Yes Yes
Whether the business is deriving 'unrealistic' income No Yes

Source: ATO table supplied to IGT 21 May 2012

4.48 As noted in chapter 2, the ATO may undertake phone reviews of taxpayers where their financial performance is outside the applicable benchmark range. Some of these phone reviews may result in a recommendation by the ATO officer that the reviewed business be the subject of a correspondence audit. Out of 8127 correspondence audits, the ATO has identified 216 which were escalated from previous phone reviews. The table below provides the results as to these case product outcomes.

Table 15: Correspondence audits escalated from phone reviews
  Number of cases Percentage of all cases
Early exit 13 6%
Nil outcome 190 88%
Outcome 13 6%
Total 216 100%

Source: ATO data supplied to IGT, 21 March 2012.

4.49 The above table demonstrates that the phone review in its current form is not a particularly effective means of identifying non-compliant taxpayers. The strike rate of these cases (6 per cent) is much lower than the general correspondence audit strike rate (24 per cent).

IGT observations

4.50 The IGT believes that the targeting of the ATO's correspondence audits may be improved by supplementing the case selection methodology with additional steps before taxpayers incur significant compliance costs, such as collating and sending records to the ATO. Whilst stakeholders presented a range of views on what other additional steps the ATO should take, the IGT is of the view that there is an important balance to be sought between the quality of the information sought and the compliance cost impact on taxpayers who may be required to supply additional information (either by phone or in writing) over and above their baseline costs.

4.51 It is instructive that many correspondence audits once commenced terminate early, after the auditor and tax agent/taxpayer discuss the issues over the phone. During that call, ATO officers are able to terminate the audit due to incorrect assumptions or details, such as the incorrect application of ATO business industry codes. It was also reported to the IGT that in certain cases audits were terminated where the taxpayer explained the benchmark variance to the satisfaction of the auditor.

4.52 The IGT recognises that the ATO needs to strike a balance between reducing compliance costs for taxpayers, whilst at the same time collecting enough information about those taxpayers to make an informed risk assessment. The IGT also recognises that there is a limit to the use of general analysis and broad metrics and that more specific taxpayer information is required in appropriate circumstances. The ATO has two options in obtaining more specific taxpayer information, namely, by way of direct contact or otherwise by third party data.

4.53 The advantage with the use of third party data is that it does not require contact with the taxpayer or any extra cost, such as completing additional forms. Thus compliance costs are minimised. In the case of benchmarking, however, once a taxpayer has been identified as being outside the benchmarks, much of the information for risk assessment (such as cash controls or record keeping practices) is accessed by direct taxpayer contact. Currently, this type of ATO activity takes place after the correspondence audit has technically started (and after the Confirmation of Audit letter is sent).

4.54 The IGT believes that such interaction should be by 'separate enquiry' prior to the correspondence audit. Making separate enquiries of the taxpayer prior to audit commencement will give the ATO an opportunity to make a more robust risk assessment (and if necessary, an opportunity for the taxpayer to directly supply additional information) enabling more accurate audit selection. Such an enquiry may be by way of a phone call or written correspondence.

4.55 ATO officers already perform these information gathering exercises, either as a phone review or as part of a correspondence audit. The IGT is of the view that auditors should undertake this information gathering consistently and separately from correspondence audits.

4.56 As indicated in Table 14 above, the IGT recognises that information about other indicators such as whether a particular business can be described as business-to-business or business-to-consumer or whether the business uses a bookkeeper may not be as easily compiled from ATO systems. Nevertheless, the IGT believes that information about these indicators should be sought from the taxpayer as part of the risk identification process prior to the commencement of any audit.

4.57 To ensure compliance costs are minimised, the IGT considers that the ATO should use a staged approach in its information gathering as part of its risk identification process. The approach should also accommodate less formal means where possible. Information gathering that occurs in a less formal context than that of a formal audit may also reduce taxpayer stress given its lower intensity.

4.58 By using a review process to gather information about the indicators of underreporting or business characteristics, the ATO would also be able to better report on indicators or characteristics that may eventually result in audit outcomes. This could be fed back into the risk identification process for future compliance selection.

Validating other identifiable indicators of underreported income

4.59 Importantly, the ATO now has completed over 7600 correspondence audit cases. In approximately 5700 cases, the ATO either found the taxpayer to be compliant or of low risk. In approximately 1800 cases, the taxpayer underreported income. This volume of cases is a potentially useful source of data which could be analysed to identify whether there are characteristics that can be used by auditors to predict either underreporting of income or full compliance. If such predictors exist, they could potentially be used early in audits to minimise not only taxpayers' compliance costs but also free up ATO resources more quickly to focus on likely non-compliance.

Recommendation 4.1

To more accurately target non-compliant taxpayers and reduce the number of compliant taxpayers being audited, the ATO should consider improvements including:

  1. examining completed correspondence cases to identify whether additional useful predictors of underreporting and/or compliance exist and use such predictors to refine the risk identification process; and
  2. implementing strategies to exclude compliant and low risk taxpayers from correspondence audits at the earliest point possible.

ATO response: Agree

The ATO will undertake a program to review completed cases with a view to refining its risk identification process. In addition it will modify the existing correspondence audit product to identify and exclude compliant and low risk taxpayers at the earliest point possible.


101 ATO case selection methodology supplied 23 February 2012.

102 ATO communication to IGT 20 March 2012.

103 The ATO considers they did not ask for responses. However, the IGT notes that approximately 25 per cent of recipients of the first tranche of this bulk mail out did respond. This could indicate that the letter was not sufficiently tested with the target market segment. The response rate for the second tranche was substantially less indicating that the ATO had made the necessary changes to the letter.

104 See also, Australian Taxation Office, The Cash Economy under the New Tax System, Australian Taxation Office, Canberra, September 2003 Page ix.

105 ibid, page 17.

106 ATO communication to IGT 23 February 2012.

107 ATO Record Keeping Audit Product Committee Minutes 25/08/2011, supplied by ATO 2 March 2012.

108 ATO case selection methodology supplied 23 February 2012.