3.1 As stated earlier, overall, stakeholders were supportive of the use of benchmarks to target non-compliant taxpayers, subject to some important reservations that are set out further below.

3.2 In their support of benchmarks, taxpayers and tax practitioners recognised that the ATO needs an effective risk identification approach and supported the need for a level playing field as taxpayers were concerned with non-compliant taxpayers gaining an unfair advantage over them. In this respect, stakeholders considered that the use of benchmarks was an appropriate approach conceptually as it could exclude from ATO compliance activities those taxpayers that were more likely to be compliant. This was generally considered to be better than conducting random audits.

3.3 Certain tax agents also advised that their advice to certain clients to improve their record keeping did not always appear to have been readily acted upon, until the taxpayer was subject to ATO compliance activity. Accordingly, these tax agents believed that appropriate awareness strategies that are cooperatively effected by the ATO in conjunction with tax agents and advisors was a helpful approach.

3.4 Stakeholders, while generally supportive of the ATO's use of benchmarks for risk identification purposes did not, however, support their use as a basis for amended assessments. Concern was also raised about the communication and delivery of the benchmark program. It was felt that the approach could have been delivered in a more effective, positive and less invasive manner. Issues of seeking to take a more cooperative approach were also raised by tax advisors in particular.

3.5 These concerns raised questions about how the benchmarks are developed, such as the accuracy and applicability of the benchmarks upon which ATO activity in this area is very much reliant. The broader issues underlying these questions relate to the transparency of a robust benchmark methodology and the data integrity of the benchmarks' inputs. Aspects of these issues are discussed below.

Transparency — Community confidence in how the ATO develops its benchmarks

3.6 All submissions to the review raised concerns about the ATO's methodology in developing the benchmarks.

3.7 Firstly, stakeholders expressed concern that there was little publicly available material that could give assurance that the ATO's methodology for developing the benchmarks was robust. In this respect, although the ATO's website publishes benchmark figures and mentions that the data is sourced from income tax returns and activity statements, it provides little extra information about the methodology for determining the benchmarks. However, it should also be noted that subsequent to submissions being received in this review, the ATO has changed its website to reveal more information about how the benchmarks are formed, and how the ATO uses them.71

3.8 Secondly, submissions asserted that the benchmarks do not account for business differences within an industry or for geographic differences. Many submissions also referred to the ATO's withdrawal of the 'cash sales benchmark' as evidence that the underlying methodology is questionable.

3.9 As a result, these submissions argued, the adverse impacts included:

  • exposing certain compliant taxpayers to unnecessary compliance costs; and
  • improperly basing amended assessments on benchmarks without specific evidence pertinent to the taxpayer.

3.10 These stakeholder concerns are discussed below.

Public assurance that the benchmark methodology is robust

3.11 In relation to providing assurance that the benchmark methodology is robust, the ATO advises that it had conducted an internal review of the benchmarking methodology in June 2011. This review was conducted by the Revenue Analysis Branch (RAB),72 being a branch within the ATO that possesses statistical expertise. This branch is also separate from the TPALS and Cash Economy operations, the area conducting the benchmarking compliance activities. The review found that 'the overall process is sound and the results are robust'73. The review also, however, recommended changes be made to:

  • improve the statistical soundness of the calculations;
  • provide more reliable data; and
  • ensure that the population is normally distributed and homogeneous and, where it is not, to either not publish the benchmark or do so with qualifications.74

3.12 The ATO implemented these changes during the course of the internal review.75

3.13 Although the ATO has conducted this recent internal review, in discussions during the review, the ATO indicated its openness to an independent third party reviewer with statistical expertise analysing the benchmark methodology. Such a review could consider whether the methodology is statistically robust.

3.14 It should be noted that the issue of whether the benchmarks themselves predict likely non-compliance is discussed in the next chapter.

IGT observations

3.15 It is important for the ATO to give public assurances about the processes used to create the benchmarks, such as publishing the methodology for developing benchmarks and the related assumptions. Such transparency would, in the IGT's view, be helpful in addressing certain stakeholders concerns and improving community confidence in the benchmarking methodology. This is especially pertinent since key benchmark ratios are used not only for audit selection, but also for taxpayer default assessments in those circumstances where the taxpayer has not provided acceptable evidence to support their reported income.

3.16 In doing so, however, there is a tension between the need for transparency, requiring the ATO to publish as much information as possible about how the benchmarks are developed and the risk that the disclosure of too much specific detail may lead to manipulation to avoid detection. Notwithstanding the recent updates to the ATO's website, the ATO could publish and better explain its methodology to allay concerns as far as it can and only reserve on those aspects that may give rise to a genuine and specific manipulation risk.

3.17 In relation to providing independent assurance that the benchmark methodology is robust, the IGT notes that New Zealand's Inland Revenue Department (IRD) has recently published industry performance benchmarks76 which were developed by Statistics New Zealand and not the IRD itself.

3.18 In this respect and to further increase confidence in the benchmarks, the ATO could seek independent assurance from ABS or other appropriate independent third party assurer with statistical expertise, that the data and methods used to develop benchmarks are robust from a statistical perspective. The ATO could then publish that assurance on its website.

Industry and business differences

3.19 As referred to above, stakeholders asserted that the benchmarks do not appropriately take into account the various differences between businesses in a given industry that may affect their financial performance, including:

  • different business models, such as online retailing versus bricks and mortar retailing;
  • different entity types, such as companies and sole traders; and
  • different business activities, such as those businesses spanning two or more industries.

3.20 One solution offered in submissions was to have a wider range of industry codes and divide some industries into smaller segments, each with their own benchmark ratios.

3.21 However, the ATO has advised that this render the benchmarks statistically invalid where the numbers in their segments are small. On its website,77 the ATO announced it had removed some smaller segments that were applicable in the 2008 year and subsumed them into a wider category for the 2009 year. For example, for the 2008 year the take away food services industry included eight separate benchmarks for the following businesses:

  • chicken shops;
  • fish and chip shops;
  • ice cream retailing;
  • kebab shops;
  • sandwich shops;
  • sushi shops;
  • takeaway pizza shops; and
  • take away food services (not otherwise captured by the above benchmarks).78

3.22 For the 2009 year, the sushi shops and sandwich shops benchmarks were subsumed into the takeaway food services industry benchmark.

3.23 The ATO advised that the reasons for subsuming the sushi shops and sandwich shops into the takeaway food services benchmarks for the 2009 year included the following:

Our analysis of the 2009 year financial data in respect of sushi and sandwich shops showed that the population size for that year was too small to pass the homogeneity and minimum population tests that all the key benchmarks must satisfy. As the sushi and sandwich shops are part of the takeaway foods industry, a decision was made to include those business types into takeaway foods.79

3.24 A comparison between the benchmark ranges for sushi shops and sandwich shops with takeaway food services for the 2008 year can be made and are set out in the table below.

Table 9: Sushi takeaways and sandwich shops — 2008 benchmarks
  Annual turnover range
Benchmark ratio Cost of goods sold / turnover Low
$75,000 to $200,000
$200,000 to $600,000
More than $600,000
Sushi takeaways 32% — 38% 36% — 44% 42% — 46%
Sandwich shops and lunch bars 39% — 53% 40% — 52% 35% — 41 %
Takeaway food services 38% — 52% 41% — 55% 37% — 51%

Source: Figures compiled from the separate 2008 benchmark pages for each of the above ATO benchmarked industries.

3.25 These benchmarks outlined in Table 9 above may also be expressed graphically (see below).

Figure 7: 2008 benchmark ranges for selected takeaway industries

Chart giving a visual representation of the 2008 benchmark ranges for selected takeaway industries.

3.26 The above table and graph show that at the low turnover range, the benchmark ranges for sushi shops have no overlap with sandwich shops or the more general category of takeaway food services. At the medium turnover range, sushi shops have some overlap with sandwich shops and takeaway food services. At the high turnover range, sushi shops have no overlap with sandwich shops, but are within the wider benchmark range of takeaway food services.

3.27 It should be noted that for a business to be selected for compliance verification a business's particular performance ratio must be a higher percentage than the applicable key benchmark.

IGT observations

3.28 Community confidence in the key benchmarks relies on the premise that the businesses within the group are sufficiently similar so that variances in financial performance would indicate underreporting of income.

3.29 Stakeholders are essentially concerned about whether they legitimately belong to a particular ATO benchmarked industry if it contained a large variety of other types of businesses. For example, a sushi shop may consider itself sufficiently different to a sandwich shop, along with all other takeaway food shops. The fact that, for 2008, separate benchmarks existed for both sushi shops and sandwich shops apart from the takeaway food services benchmarks may indicate that these businesses are distinct enough to have different ranges of financial performance.

3.30 The IGT recognises that whilst businesses would have greater confidence in the benchmarks if they were more specific to their industry, there are also limitations to the level of specificity.

3.31 In the example above, the ATO determined that the population sizes for sushi shops and sandwich shops for the 2009 year were too small to be maintained as separate ATO benchmarked industries without affecting the statistical validity of the benchmarks. In deciding to combine these businesses into the takeaway food services ATO benchmarked industry, the ATO relied on the fact that 'sushi shops and sandwich shops are part of the takeaway foods industry'.

3.32 Whilst this may be true at a broad level, there is a risk in making assumptions that make the comparison of the businesses' financial performance less likely to indicate non-compliance and, therefore, less likely to identify potential underreporting.

3.33 This risk could be minimised by the ATO undertaking analysis to satisfy itself and taxpayers that sushi shops and sandwich shops, at a statistical level, belong to the takeaway food services ATO benchmarked industry. Publication of this analysis would also help to engender community confidence in the benchmarks.

3.34 The above example also implies that different businesses within a broader industry classification will 'cluster' their financial performance at different points along the broader industry's benchmark range. One implication is that certain industries may tend to be selected for benchmarking compliance activity more by reason of commercial factors rather than potential underreporting.

3.35 The ATO has sought to mitigate this potential problem by widening the benchmark range or increasing the variance threshold before a case is selected for compliance verification.

3.36 In the IGT's view, however, where the statistical analysis indicates that these smaller ATO benchmarked industries are significantly different to the broader ATO benchmark industry into which they are intending to be combined, they should be excluded from that broader ATO benchmark industry. To better identify potential non-compliant taxpayers in these smaller industries alternative strategies could be considered rather than employing benchmarks as a risk identification tool in isolation. The following chart seeks to illustrate the circumstances in which combining an industry with another may or may not be appropriate.

Figure 8: Combining ATO benchmarked industries

Flow chart regarding combining ATO benchmarked industries.

View image enlarged

Geographic differences

3.37 A consistent concern raised in submissions was whether benchmarks appropriately take into account geographic differences. Examples included the difference between businesses located in different states and territories, between metropolitan and country businesses, between shopping malls and on streets, certain city suburbs and the central business district itself.

3.38 These differences were thought to materially affect financial performance of businesses, in particular the effect of differing freight and labour costs, and may also indicate seasonal issues, such as those experienced in holiday towns or in fruit picking districts.

3.39 To help understand the issues, stakeholders have suggested that the ATO should publish benchmarks based on states or localities or at least show why they have not done so.

3.40 In relation to the difference of cost of sales average ratios between metropolitan and country businesses, the ATO advises that the difference is only three percentage points (see Table 10 below).80 This analysis was conducted by assigning the population used to develop the small business performance benchmarks to either country or metro based on their business postcode and a list of reference postcodes. Following the same methodology used to develop the performance benchmarks, the ATO then calculated cost of sales ratios for each entity and thereafter assigned each to a turnover range. From these figures, averages (that is the mean) were then calculated.

3.41 For all industries combined, the ATO provided the following table comparing cost of sales ratios averages (that is the mean).

Table 10: ATO country and metro cost of sales ratios for the 2009 year — comparison of means
Turnover range Country (mean) Metro (mean) Difference
Low turnover range 44% 41% 3%
Medium turnover range 50% 46% 3%*
High turnover range 55% 51% 3%*

Source: ATO communication to IGT, 31 January 2012.

* the discrepancy of 1% for each is presumed due to rounding.

3.42 The ATO also provided a comparison for the restaurant industry segment as identified (see Table 11 below).

Table 11: ATO comparison of Country and Metro restaurant industry cost of sales ratios for the 2009 year
Turnover range Country (mean) Metro (mean) Min Benchmark Max Benchmark
Low 37.10% 35.83% 32.81% 39.19%
Medium 37.11% 37.47% 34.10% 39.90%
High 32.87% 34.58% 30.99% 36.93%

Source: ATO communication to IGT, 31 January 2012.

3.43 Table 10 above shows that, broadly for both country and metro businesses in all benchmarked industries, there are some differences in the average cost of sales ratios — three percentage points. With respect to the restaurant industry in particular, Table 11 above shows there are some minor differences in the average cost of sales ratios as between country and metro restaurant businesses.

3.44 Importantly, in the case of the restaurant industry, the benchmark range used by the ATO (as indicated by the Min Benchmark and Max Benchmark columns above) seeks to accommodate those differences.

3.45 The ATO has also advised that it seeks to adjust the cost of sales figures data obtained from tax returns. However, labour cost amounts are the only adjustment made, removing it from the cost of sales figures. Therefore amounts such as freight would remain in the cost of sales figure.81 The question remains as to whether geographic differences, such as the increased cost of freight, affect the accuracy of the benchmarks materially.

IGT observations

3.46 Based on the ATO material outlined in Table 11 the differences in the average of cost of sales ratios between country and metro restaurant businesses may not be as wide as they may have been perceived. Suggested reasons for these similar average performance ratios include that higher labour and freight costs are passed on in the form of higher prices.

3.47 This analysis, although a national average, indicates that the ATO may foster greater community confidence in the benchmarks by publishing this type of information for other industries.

3.48 The ATO has not, to date, conducted any similar analysis with respect to differences between states and territories. For example, submissions expressed the view that states experiencing a mining boom may mean businesses are operating under substantially different conditions compared to those states that are not experiencing a mining boom. It was also suggested that states with particular remoteness or transport access issues may also have important differences that need to be considered. To allay stakeholders' concerns, it would be beneficial if the ATO conducted and published such an analysis similar to that done for the country/metro comparison.

3.49 It should be noted, however, that further analysis to distinguish between specific geographical differences may be difficult to accomplish consistently without imposing greater costs.

Cash sales benchmark

3.50 As referred to above, submissions also pointed to the ATO's withdrawal of the cash sales benchmark as a reason to question the methodology of the performance benchmarks.

3.51 The history of the cash sales benchmark is very unfortunate for those taxpayers who were inadvertently affected, particular where they were subject to ATO default assessment, although these assessments were later reversed.

3.52 The ATO advises that the cash sales benchmark relied on a different data source (that is the financial institutions' records) and methodology to that of the financial performance benchmarks. The reasons for the ATO's withdrawal were inconsistencies in the way in which the financial institutions record cash-outs paid by businesses to their customers. By way of comparison, the performance benchmarks rely on information reporting in taxpayers' income tax returns. This indicates that community confidence in the benchmarks also relies on the integrity of the data inputs, which is an issue discussed further below.

3.53 In the light of the above discussion on the transparency of aspects of benchmarks, the IGT makes the following recommendation.

Recommendation 3.1

The ATO should improve community understanding and confidence in the ATO's industry performance benchmarks by publishing material that seeks to provide assurance that the methodology to develop the benchmarks is robust and communicating it more broadly. Such material should include:

  1. all current benchmarking inputs and methodology (except for specific aspects giving rise to genuine manipulation concerns);
  2. assurance from an independent party with statistical expertise about the robustness of the benchmarking methodology;
  3. assurance that the types of businesses included in ATO benchmarked industries are sufficiently similar to each other to compare financial performance; and
  4. appropriate comparisons and explanations of ratios between
    1. country and metro businesses within each benchmark; and
    2. states and territories on an aggregated basis.

ATO response: Partially agree

The ATO agrees with sub-recommendations 3.1(a), (b) and (c) and disagrees with sub-recommendation 3.1(d).

We agree to enhance our existing communication and materials about benchmarks to provide more detailed information about how benchmarks are developed.

We will seek assurance from an independent party with statistical expertise on the robustness of the benchmarking methodology and will publish this, along with assurance on the similarity of business in benchmarked industries, and further information about the inputs and methodology for benchmarks.

In relation to sub-recommendation 3.1(d), we appreciate the underlying intent of this recommendation and are committed to providing information which will improve understanding and confidence the community has in the industry benchmarks.

However, we are concerned that publishing large numbers of additional country and metro benchmarks, or benchmarks for comparison between states and territories on an aggregated basis, will not assist and has the potential to cause confusion.

We note that your report includes tables following paragraph 3.41, which reflect that any differences associated with locality are not significant (there being less than 3% difference between country and metro averages).

Our experience shows that while prices of goods and services may be higher in certain states, territories and regions, the costs of the inputs to those goods or services are also generally higher. This results in the relatively small variation noted above.

3.54 As referred to above, stakeholders raised concerns with aspects of the data integrity for benchmark inputs. These are discussed below.

Data integrity — correct industry allocation identification

3.55 Stakeholders raised concerns about the process the ATO uses to allocate businesses to ATO benchmarked industries.

3.56 After cases are selected and allocated to an ATO officer, the officer conducts some initial internal research, such as checking ATO systems or the business's website, to determine whether the business is still operating or the correct benchmark has been applied. The ATO officer may end the correspondence audit before contacting the taxpayer or sending the audit confirmation letter where the ATO officer determines that the business:

  • has ceased trading;
  • actually belongs to a different ATO benchmarked industry, and the business's ratios now report within the benchmarks for that correct industry; or
  • does not belong to the ATO benchmarked industry, and does not appear to belong to any other ATO benchmarked industry.

3.57 As previously stated, these cases referred to above are recorded as an 'early exit' by ATO audit teams.

3.58 The ATO has also advised that are there instances where an auditor may be satisfied that the correct benchmark was applied and commences the audit but the audit may still be closed 'early'. This may occur where upon contacting the taxpayer, the auditor may agree with the taxpayer that one of the three circumstances described above are satisfied and closes the audit. Since the taxpayer has already been contacted, however, it is not reported as an 'early exit', but rather as a 'nil outcome' by the ATO.

3.59 A 'nil outcome' also denotes a case in which the ATO has verified that the taxpayer has appropriate records and all income has been reported.

3.60 The following table indicates the proportion of initially selected cases from January 2010 (when benchmark strategy compliance activities began) until April 2012 which were closed as 'early exit'.

Table 12: Early exit rates
Intensity Product Early exit Total initiated cases Early exit rate (a)
Highest (field) Cash Economy Audit 65 341 19%
(field) Specific Audit 7 150 4%
(desk) Correspondence Audit 457 8127 6%
(field+ desk) Record keeping audit 526 1453 36%
(desk) Phone review 374 10213 <1%

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

Note (a): rounded to the nearest per cent.

IGT observations

3.61 The early exit rates give an initial indication of the integrity of the data used to create the above compliance activities. This integrity is subsequently confirmed or contradicted by profiling work which the auditor undertakes before contacting the taxpayer or otherwise terminating the compliance activity.

3.62 The IGT supports the process of the auditor conducting internal research before contacting the taxpayer to commence the audit. It is important that automatic filters and processes, such as those used to allocate businesses to ATO benchmarked industries, be supported by audit officer analysis and interpretation. It is appropriate that auditors be satisfied that the audit ought to proceed, and not commence the audit simply on the basis that the case was allocated to them.82

3.63 With respect to correspondence audits, an early exit rate of 6 per cent indicates that the case selection methodology is 'initially supported' in that the limited profiling of auditors only closed a small proportion of cases prior to commencement. For a more accurate profile, auditors would normally need to contact the taxpayer or tax agent for business specific information.

3.64 As noted above, the ATO benchmarked industry allocation process partially relies on the use of ATO business industry codes. Whilst the preceding section indicates that this process is initially supported by the low early exit rates, stakeholders raised specific concerns with respect to the use of ATO business industry codes in this process. These concerns are addressed below. It should be noted that, in their submissions, stakeholders often used the term 'ANZSIC codes' to refer to the ATO business industry codes.

Data integrity — limitations with industry codes and errors in identifying correct codes

3.65 The ATO benchmarking process categorises taxpayers into certain ATO benchmarked industries based on the ATO business industry codes (a 5-digit code derived from the ABS's 4-digit ANZSIC codes) and other tax return data.

3.66 In completing tax returns, taxpayers and their representatives often rely on the ATO's publication Business industry codes to locate their correct code. These ATO codes are presented in a web-based (that is PDF) booklet in two lists:

  • alphabetically, independent of hierarchy or grouping for example 'Abalone Fishing' is followed immediately by 'Abattoir operation — except poultry'; and
  • hierarchically, with businesses grouped under broader industry categories.

3.67 The introduction to the ATO booklet indicates:

Only use codes from Business industry codes 2012, and take care to describe your business activity as accurately as possible. An incorrect code may result in clients not receiving necessary service or material from the ATO, or could lead to incorrect targeting of compliance activities.83

3.68 Stakeholders have advised the IGT that the codes and industries supplied in the Business industry codes booklet have certain inadequacies. Common complaints in the submissions are:

  • Variations in business models are not accounted for. For example, wholesale versus retail or business-to-business versus business-to-consumer.
  • The method of searching for the correct industry in itself is not intuitive. There is no ability to 'drill down' into categories or there is insufficient explanation or linking or cross referencing to other potential categories.
  • Specific subtypes of industries that may logically fall under a wider category but have very different cost structures are not accounted for. For example, a commercial painter is considered a 'painter' and thus treated the same a domestic painter. Those in the industry consider commercial and domestic painters to be completely different businesses.
  • The 'not elsewhere classified' (NEC)84 option is not included for some industries, so businesses are forced to choose something that is not precise.
  • Certain businesses are not identified at all. For example, home theatre installers.
  • Mixed businesses are not accommodated.
  • Although the name of the industry may be correct in a general sense, the description attached to the industry does not actually reflect the reality of the taxpayers' business or there are other specific factors that render the classification inaccurate.

3.69 Industry names themselves may be oddly worded and do not reflect plain language usage. Furthermore, taxpayers' and even some tax agents' do not have a strong awareness of the codes and their use for potential or actual audit by the ATO. This, combined with the difficulties above and pressures to minimise compliance costs, has resulted in taxpayers and tax agents:

  • choosing a code just because they need to enter 'any value' in that label in order to continue completing the tax return;
  • rolling over the code from last year's tax return without much thought on additional action; and
  • 'wasting' significant time in searching for a code that may or may not be more accurate.

3.70 As a result of this general lack of awareness, concerns were raised by other stakeholders that:

  • benchmark data is therefore not truly reflective of the actual population; and
  • certain taxpayers are being incorrectly selected.

3.71 The ATO advised that businesses are not allocated to an ATO benchmarked industry on the basis of the ATO business industry code alone. Rather, some data adjustments are undertaken. That is, a business must also have certain keywords present in its trading name and/or business description (from the income tax return) to qualify for the relevant benchmarked industry.85 For example, if a book retailer (code 42442) incorrectly chose the code 42441 (newsagents), it would not be allocated to the newsagency ATO benchmarked industry unless the book retailer also had 'newsagency' related keywords in its trading name or business description.86

3.72 During discussions with the IGT, the ATO advised that ATO staff have internal access to an ABS-developed and maintained custom software program, called the 'ANZSIC Coder'87 referred to in this report as the 5-digit Coder. This 5-digit Coder allows staff to:

  • enter a keyword, and search for possible 5-digit ATO business industry code matches;
  • enter all or part of a 5-digit ATO business industry code and see what industries it relates to;
  • view a hierarchy of industries and drill-down to more specific industries; and
  • view the description and primary activities for a given 5-digit ATO business industry code.

3.73 ATO staff can use this program to help ascertain whether the taxpayer has used the correct ATO business industry code. For example, an ATO officer may use the coder whilst reviewing the business description on an income tax return or while having a phone conversation with a taxpayer or their tax agent about their business activities.

3.74 The ATO has advised that, since 2002, the 5-digit Coder and associated data files were made available to developers of Electronic Lodgment Service (ELS) software. It indicated that 'some ELS software producers have included the coder in their software while others have built similar functionality into their software' using the associated 5-digit Coder data files.88

IGT observations

3.75 The ATO's benchmark development process does not solely rely on the ATO business industry code selected by the taxpayer or their tax agent in their income tax returns as some data adjustments are undertaken. This measure mitigates the risk of incorrect industry allocation to a certain degree. The IGT is of the view, however, that the ATO can strengthen data integrity by making it easier for taxpayers and tax agents to select the correct ATO business industry code in the first instance.

3.76 The IGT considers that the ATO may better support more accurate code identification and entry in the Australian Business Number application form and tax returns in a number of ways.

3.77 First, the ATO could make it easier for taxpayers and tax agents to find the right code, for example, by modernising and publicly releasing the ABS-developed 5-digit Coder described above. The 5-digit Coder is currently used within the ATO as an installable program. Widespread adoption may be more likely if the coder is web-based, therefore not requiring the user to download any program. Web-based versions also ensure users have access to the most up to date database. Whilst the ABS itself currently has a web-based search facility for its 4-digit ANZSIC codes,89 it does not have the same information or functionality as the 5-digit Coder used within the ATO.

3.78 Second, the ATO could give taxpayers and tax agents the ability to find the right code through either browsing (that is drilling down through a hierarchy) or searching. The 5-digit Coder currently used in the ATO has both a browse and search function.

3.79 Third, the ATO could better communicate with taxpayers and tax agents at the time of lodgment about the importance of improving ATO code identification. For example, electronic lodgment programs could provide a prompt at the Business industry code field to remind users about the importance of the code and providing a link to the coder.

3.80 Fourth, the ATO may consider adding 'primary activities' to existing ATO Business industry codes for those ATO benchmarked industries that are currently accounted for under a residual category within the code. For example, whilst kebab shops are their own ATO benchmarked industry, they are not mentioned in the Business industry codes. Rather, they would belong to code 45120 by virtue of the residual category 'Take away foods — retailing'. By specifically mentioning them, kebab shops may join the already listed pizza shops and hamburger shops as businesses that can readily identify themselves as belonging to the shared code 45120. This measure, however, does not assist the ATO to determine if the business is a kebab shop. The code only indicates to the ATO that the business is a takeaway food retailer. In this case, the ATO would continue to rely on keywords in the trading name and business description in the income tax return to help it identify the business as a kebab shop. The intent is to assist the taxpayer or tax agent to quickly select the correct code, even if it is currently general in nature.

3.81 Fifth, where the current ATO Business industry code system does not provide the current level of granularity required for benchmarking purposes, the ATO may consider increasing the use of the fifth digit in the ATO Business industry codes to provide it. To illustrate this point, ATO Business industry code 41290 is currently shared by 12 specific food retailers. An option may be to create, for example, code 41291 to cover bread shops and bakeries, code 41292 for delicatessens and small goods retailers, code 41293 for cake shops and patisseries. That is, there is potential to customise the fifth digit to create a unique code for each ATO benchmarked industry.90 This may also provide a useful opportunity to enhance the data integrity or the information collected in benchmarking for future compliance activities.

3.82 Although greater granularity could be provided, it is important to note that such a task may be a significant undertaking and that such action should be commensurate with the aims of maintaining reasonable consistency over time to minimise user and comparative difficulties. Such an undertaking may necessarily require consultation with Government and the ABS.

Recommendation 3.2

To assist taxpayers and tax agents in identifying the correct ATO business industry codes, the ATO should take the necessary steps (and where necessary, work with the ABS) to provide public access to the ABS developed 5-digit Coder.

ATO response: Agree

The ATO will consult with the ABS and take steps to make the 5-digit coder publicly available within the capacity of the forward information technology program of the ATO.

Recommendation 3.3

The ATO should improve its support for tax agents and taxpayers in helping them to include the correct ATO business industry code in the tax return form, such as by:

  1. enhancing communication regarding the importance of correct industry code selection and the impacts it may have in subsequent ATO compliance activities; and
  2. assuming the 5-digit Coder is made publicly accessible, seeking external user feedback on the coder's operation to identify improvements required to make it more effective, including the code structure used for the fifth digit.

ATO response: Agree

The ATO will enhance our communication about the importance of correct industry code reporting on tax returns through existing channels, including 'Tax Time' briefings and support materials.

The ABS has responsibility for the ANZSIC code structure which is subject to an international framework (4 digits). They have an existing public consultative mechanism, which includes linking new business descriptions to existing industry groupings. The ATO will seek feedback from key stakeholders, including industry representatives, when adding or removing a new sub-group at the 5th digit level.

Data integrity — limitations with industry codes for mixed businesses

3.83 Stakeholders raised concerns that since mixed businesses are only accounted for by one industry code on the income tax return, any benchmark is unlikely to apply accurately. Stakeholders were concerned, therefore, that mixed businesses faced the risk of being incorrectly chosen for a correspondence audit, or even the prospect of having a default assessment issued against them on the basis of an incorrect benchmark.

3.84 Certain tax agents reported that they could satisfy the ATO by disaggregating their clients' financial reports into their separate business lines and calculating individual ratios for each of the 'industries' and show how it falls within the benchmarks. However, it is a costly and time consuming exercise. Further, the IGT was advised that generally small businesses of this nature do not disaggregate their different activities as part of their normal record keeping (that is such a process is not aligned with natural business systems).

3.85 With the exception of mixed primary/non-primary production activities, the existing income tax return form and related schedules do not provide for any distinction between the business lines of a mixed business. The ATO has advised, however, that several income tax return schedules contain labels that allow a taxpayer to indicate that it is a mixed business.91 Where the taxpayer has done so, the ATO will not use that taxpayer for calculating industry benchmarks, nor will the ATO select that taxpayer for benchmark related compliance activity.

3.86 The ATO also advises that where the mixed business label has not been completed, then the taxpayer may be selected for compliance activity. In the case of a correspondence audit, the auditor will often identify whether the taxpayer is a mixed business during the phone call that occurs after the audit confirmation letter is sent. During that phone call, the auditor will discuss the nature of the business and the taxpayer's record keeping practices. The auditor may decide to close the audit early if they determine there is a low risk.

3.87 Alternatively, where the auditor has reason to believe that the taxpayer has not correctly recorded all of their income, they may proceed with an audit as planned. In such cases where the audit proceeds, if the auditor determines there has been omitted income, the auditor will seek to use disaggregated sales or cost of sales figures to apply the relevant benchmarks on a weighted basis. Where disaggregated figures are not available, the auditor will ask the business owner to provide purchase records and the owner is to 'apportion these expenses between each enterprise as it believes would be appropriate and provide this information to the auditor.'92

3.88 The ATO has advised the IGT that before, and at the time an interim report is issued, taxpayers may supply extra information for example:

The taxpayer can provide a statement and evidence where they are a mixed supply and the relevant benchmark does not apply, for example mixed food retail/take away businesses.93

3.89 The ATO has also advised94 that:

The methodology we used to select businesses that fall into industry categories seeks to exclude those that identify themselves as mixed businesses.

When we carry out an audit activity the ATO takes into account the personal circumstances of the business we are reviewing. If it is identified as a mixed business, the ATO may apply the benchmarks proportionally to the separate business types.

3.90 By way of example, one Siebel case took 288 working days and resulted in a complaint by the taxpayer/tax agent after an interim report was issued. The auditor did attempt to disaggregate the business, and applied two different benchmarks to two different parts of the business. The auditor was not satisfied with the quality of the records, and thus proposed default assessments. Once the taxpayer lodged a complaint, the case was reviewed by another officer. The case was subsequently closed with no further action on the basis that the business actually had four distinct business lines (rural petrol station, tyre shop, hardware shop, agricultural supplies shop), of which only one business line was outside the benchmark. For that one business line, it was determined that particular agricultural purchases did not fit in that benchmark, and that once removed, the business line fell into the benchmark. The reviewing officer also came to a different conclusion about the risk posed by the quality of the records. Importantly, the reviewing officer also concluded that the 'taxpayers business activity does not fit within the parameters of the benchmark'.

IGT observations

3.91 The benchmarking approach is generally premised on comparing the financial performance of like businesses. In the circumstances of mixed businesses, this may be difficult to do. Therefore, an attempt to further refine the relative risks of mixed businesses' non-compliance should not be based on a comparison of financial performance with other businesses.

3.92 By using the mixed business indicators on tax returns to exclude mixed businesses from the calculation of benchmarks, the ATO has recognised the limitations of the benchmarks as currently implemented. The IGT welcomes the ATO approach.

3.93 Nevertheless, the IGT is concerned that a mixed business can be the subject of an audit by reason of not completing the mixed business label on the tax return. If they had done so neither the taxpayer nor the ATO would have needed to consider a benchmarking originated audit.

3.94 Stakeholders have cited examples of disaggregating business lines to show the ATO each business line is within the applicable benchmarks so as to close the audit early. The IGT is of the view that businesses and their advisors would usually take this course of action if they perceive that it would cheaper to do so than allowing the audit to take its normal course by supplying the ATO with the requested records.

3.95 Stakeholders also cited an example of using disaggregation to refute a default assessment based on the benchmark.

3.96 These are difficulties that are faced by taxpayers and auditors alike in disaggregating mixed business. The IGT is of the view that it would be generally impractical for the ATO to require small businesses to disaggregate business lines for the simple purpose of testing whether they are within the benchmarks. To do so would be to impose unnecessary compliance costs, especially since the ATO may be able to take into account risk factors other than benchmarks.

3.97 Whilst it may be difficult to benchmark mixed businesses, they, like any other affected business, may be subject to ATO review for adequacy of record keeping or underreporting income or seeking excessive deductions. It is therefore important that the ATO consult with the community on how to best engage with mixed business owners.

3.98 The ATO benchmarking strategy is generally useful for comparing like businesses' financial performance. Mixed businesses, however, are typically difficult to compare. In addition to the benchmarks, the ATO currently uses other models and approaches to address the cash economy, such as the cash economy risk model and data matching projects. The IGT believes that there is scope for the ATO to consider other risk models to address cash economy mixed businesses in a manner that minimises compliance costs.

Recommendation 3.4

To minimise compliance costs and improve mixed businesses risk identification, the ATO should consult with tax agents and taxpayers with a view to developing and employing alternative risk models and approaches to assess the risk of underreporting by mixed businesses.

ATO response: Agree

The ATO will consult with tax agents and taxpayers with a view to reducing compliance costs for mixed businesses and improve risk identification for those businesses.

Data integrity — businesses 'code' their expenses differently in tax returns, affecting their own benchmarks

3.99 Concerns were raised that some taxpayers were outside of the benchmarks (usually the cost of sales benchmark) because of the way they recorded their expenses and what they included in cost of sales. Cost of sales is a field on both the Company Tax Return (Question 6 Expenses, label A) and the Business and Professional Items Schedule for Individuals (P8 Label KLM). For example, some include in their cost of sales:

  • only the cost of stock;
  • cost of labour;
  • cost of rent;
  • capital acquisitions; and
  • lottery sales minus commissions (for example newsagents), rather than only adding sales commissions to business income on a net basis.

3.100 One tax practitioner representative body advocated for the ATO to provide more education about common errors that tax practitioners make, in particular, the inclusion of labour in cost of sales. In this respect, the ATO has advised that it intends to develop a fact sheet to assist tax return preparers to correctly calculate the cost of sales figures.95

3.101 In terms of calculating the benchmarks, the ATO states that it takes into account whether wages are incorporated in the cost of sales figures. Where a taxpayer has indicated wages in another label on the income tax return form and has indicated that the wages are part of cost of sales, the ATO will deduct the wages figure to reduce the cost of sales figure.96

3.102 The ATO has advised that it also makes this same subtraction (where applicable) when calculating default assessments based on the benchmark.97

IGT observations

3.103 Whilst the practice of including labour in the cost of sales varies, it is not necessarily incorrect to do so. Where labour is directly attributable to the acquisition or conversion of inventory, then it is appropriately included as part of cost of sales. For example, a bicycle shop that buys parts and assembles bicycles before selling them would include labour in their cost of sales.98

3.104 With respect to cost of sales, therefore, the IGT appreciates that the ATO practice of subtracting labour from the cost of sales figures ensures that the data is not affected by the practice of including or excluding labour from cost of sales. This assumes, however, that return preparers are correctly entering the details in the return and schedules.

3.105 The issue of taxpayers correctly entering their figures and details in the tax return is further complicated when taxpayers erroneously include other amounts in the cost of sales figures. These amounts include capital costs or other overheads, and expenses not directly related to the acquisition, conversion or manufacture of goods. Once these incorrect figures have been entered into the tax return and lodged, there is little the ATO can do to detect this error.

3.106 The responsibility to correctly complete a tax return is that of the taxpayer. The taxpayer's agent or representative has certain duties and responsibilities too. Nevertheless, the ATO is in a position to assist taxpayers and tax agents in correctly entering cost of sales amounts. The optimal means to do this could be through a systems change, so that when costs of sales is entered on a field on an electronic tax return form, a prompt is triggered to remind the tax practitioner what types of items to include or exclude in this field. Such a change, however, may involve complexities as alterations of the ATO's information technology upgrade schedule may be limited. As an alternative measure, the ATO could assist through a targeted communications strategy.

3.107 The ATO could also make it easier for small businesses to correctly complete the return and schedules by tailoring the returns to their needs.

Recommendation 3.5

The ATO should seek to continue to improve the data integrity of benchmark inputs by, for example, better assisting taxpayers and tax agents to include the correct items of costs in the cost of sales label on the income tax return.

ATO response: Agree

The ATO will enhance our communications about benchmarks to assist taxpayers and tax agents to include the correct items of costs in the cost of sales label on the income tax return by improving our website material and advice to tax agents.

Bulk mail out letters program — impacts

3.108 Stakeholders raised concerns about the first bulk mail out of advisory letters to taxpayers and tax agents regarding the benchmarks. As mentioned above in Table 4, 37,847 businesses (and 20,967 of their tax agents) were sent letters in 2010 advising them that they were outside the small business benchmarks. A sample of the first batch of letters is included in Appendix 7.

3.109 These letters included instructions such as:

If you think you may have made an error or left something out, we have enclosed a form that will help us correct anything that might need to be corrected. ...

If you are satisfied that you are meeting your tax obligations correctly, you do not need to contact us or do anything further.

3.110 The letter indicates that the taxpayer or tax agent need not respond to the ATO if they were satisfied that there were no errors. Nevertheless, a significant proportion of letter recipients (9796 or 25.8 per cent) wrote a letter to the ATO for no other reason than to provide an explanation of why they were justifiably outside the benchmarks. Submissions advised the IGT that a proportion of taxpayers wrote to the ATO directly while others asked their tax agent to do so. Certain tax agents charged their clients to write these letters on their behalf, while others indicated that they did not or could not charge their clients, despite each letter taking significant time to complete. In either case, a compliance cost is borne either by the taxpayer in the form of fees, or by the tax agent in the form of uncharged work.

3.111 The original enclosed form issued in May 2010 letters, called the Correcting Tax Errors form, was designed only to allow the taxpayer to notify the ATO of any errors the taxpayers found when reviewing their business records. Subsequent versions of the form contained an extra tick box stating:

No, I do not need to make a correction. However, I would like to provide an explanation.

3.112 Despite this modification, none of the advisory letters themselves indicated when or why the taxpayer ought to provide such an explanation. While the addition of this tick box may indicate that the ATO was expecting explanations from business owners, the letter did not provide a prompt to do so.

3.113 Furthermore, stakeholders also observed that after sending in an explanation (rather than a notification of an error) they did not receive any acknowledgement from the ATO and were thus left in a position of uncertainty about whether the ATO was satisfied with their explanation. One representative group also expressed concerns that the explanations given were not recorded by the ATO since some of their clients were later chosen for an audit. When the tax agent asked why the ATO did not consider the explanation previously given to the advisory letter, the auditor did not have a record of it.

3.114 Stakeholders further submitted that the tone of the letters was accusatory. Tax practitioners indicated that many of their clients experienced stress as a result or were unsure of why they were receiving the letters. Tax practitioners also raised concerns that some of their clients attributed the receipt of these advisory letters to their agent's mismanagement, with some tax practitioners relaying the business's sentiment that they 'wouldn't have attracted the attention of the ATO if their agent was doing their job properly'.

3.115 Some tax agents expressed frustration with the fact that the letters were sent directly to taxpayers, even where the tax agent was listed as the authorised contact. Tax agents submitted that they received phone calls from clients regarding benchmarking letters which the tax agent knew nothing about. Tax agents felt this 'put them on the back foot', and expressed concern that it made them look unprofessional and undermined their relationship with their client. One representative body expressed the view that if the tax agent is the authorised contact, then all communication should go to them first, and it is their prerogative rather than the ATO's, to determine how to approach the client. Nevertheless, some tax agents also expressed the view that they did not mind having the ATO contact their clients directly for these types of matters, as long as the tax agent was given the choice or at least informed first.

3.116 The ATO has made changes99 to their benchmark advisory letters stemming from feedback and complaints received from letter recipients, and feedback from ATO consultative groups. The ATO advises that these changes include:

  • where a tax agent is listed, the ATO will send a letter to them first, indicating that the ATO will send a benchmark advisory letter to their client between 5 days to 2 weeks later;
  • where a taxpayer or tax agent provides a response to the ATO (be it a notification of an error or an explanation), the ATO will send an acknowledgement letter to them (in June 2011, the ATO sent 17,000 acknowledgement letters to those who responded during the period 1 July 2010 to March 2011);
  • since May 2011, all cash economy letters are user tested in the ATO's Simulation Centre to address the issue of tone, content and readability of the letters;
  • engaging an external consultant to provide feedback on some cash economy letters (that is data matching letters) and incorporating that feedback into all their other cash economy letters; and
  • where an explanation is provided in response to an advisory letter, the ATO now scans this response and attaches it to the taxpayers' file in Siebel, so that the response is now considered by the ATO before deciding on potential compliance action.100

3.117 These updated letters were sent in November and December 2011. A copy of this letter to taxpayers can also be found in Appendix 7. The letter provides more detail about record keeping requirements and where taxpayers can obtain assistance. It also does not include an enclosed form for notifying errors.

IGT observations

3.118 The IGT is of the view that many of the issues raised by stakeholders in relation to the first bulk mail out letters program can be attributed to the original design of the letters themselves and the process of issuing the letters and handling the responses. The ATO did not, at that time, have a process in place to adequately record the explanatory responses to its letters.

3.119 The ATO's practice of contacting the taxpayer directly where the tax agent was an authorised contact appeared to be based on the assumption that, in relation to the matters raised by the letter, the taxpayer rather than the tax agent was best placed to action the letter. Whilst this may be true, the ATO did not fully appreciate the practical importance taxpayers place on their tax agent in dealing with the ATO.

3.120 The IGT considers that given the important role that tax agents play in this particular market segment (some 95 per cent of taxpayers are represented), it is important that the ATO liaise directly with tax agents to ensure they are appropriately informed of any actions of this kind in the future.

3.121 The IGT acknowledges the iterative improvements the ATO has made in relation to its letters program, in particular, the second bulk mail out letters program being based on the feedback provided. The introduction of user testing will also be beneficial in addressing design issues before letters are issued in such a manner.

3.122 It remains to be seen whether the above recent improvements address the concerns of taxpayers and tax agents. However, in future, the ATO should consider preceding such 'bulk' programs with a pilot and/or consultation with relevant stakeholders.

71 Australian Taxation Office, Small business benchmarks - overview, Australian Taxation Office, Canberra, 21 February 2012.

72 ATO communication to IGT, 31 January 2012.

73 ATO internal minute from RAB to Cash Economy Risk and Strategy dated 27 June 2011 supplied by ATO 31 January 2012.

74 ibid.

75 ibid.

76 New Zealand Inland Revenue (IRD), Industry benchmarks: About industry benchmarks, IRD, Wellington, 19 April 2012, viewed 2 July 2012.

77 Australian Taxation Office, Cash economy frequently asked questions: Cash economy phone review, Australian Taxation Office, Canberra, 21 February 2012, viewed 23 February 2012.

78 Australian Taxation Office, Takeaway food services - issued 2010, Australian Taxation Office, Canberra, 7 October 2011, viewed 2 July 2012.

79 ATO communication to IGT, 20 March 2012.

80 ATO communication to IGT, 31 January 2012.

81 ATO communication to IGT, 20 March 2012.

82 Correspondence audit procedures - supplied by ATO, 27 January 2012.

83 Australian Taxation Office, ATO Business industry codes, Australian Taxation Office, Canberra, 30 May 2012, viewed 2 July 2012, NAT 1827-6.2012.

84 Some industries categories are broken down into more specific subtypes, and where the business does not fall into these subtypes, they usually have the option of choosing 'NEC'.

85 ATO communication to IGT, 31 January 2012.

86 ATO communication to IGT, 31 January 2012.

87 Australian Taxation Office, Industry coding of business returns - August 2002, Australian Taxation Office, Canberra, 28 March 2006, viewed 2 July 2012.

88 ibid; ATO communication to IGT, 30 March 2012.

89 Australian Bureau of Statistics, ANZSIC 2006 Search facility, Australian Bureau of Statistics, Canberra, 25 September 2008, viewed 6 June 2012.

90 See Appendix 4 for an explanation of how this fifth digit works, and how it relates to the ABS four digit system.

91 Business and Professional Items for Individuals: P3 Label B; Company Tax Return: 3 Label E1; Partnership Tax Return and Trust Tax Return: 2 Label B1.

92 Australian Taxation Office, Small business performance benchmarks default assessment calculation method - mixed enterprises, supplied by ATO, 31 May 2012.

93 ATO communication to IGT, 23 January 2012.

94 ATO communication to IGT, 31 January 2012.

95 ATO communication to IGT, 31 May 2012.

96 Income tax labels used to calculate benchmarks, supplied by ATO, 23 February 2012.

97 ATO communication to IGT 23 January 2012.

98 West Australian Small Business Development Corporation (SBDC), Components of a profit and loss statement, SBDC, Perth, 2012, viewed 2 July 2012.

99 ATO communication to IGT 23 December 2011.

100 ATO communication to IGT, 30–31 January 2012.