ATO processes to cleanse and refine data

2.1 A number of submissions have raised concerns that the ATO does not appear to be sufficiently refining and cleansing data before comparing it against taxpayer information to identify any discrepancies. Submissions frequently singled out CGT on real property disposals as an example.

2.2 Specifically, stakeholders were concerned that the ATO did not seem to appreciate the differences in timing between execution of the contract and the property settlement date in relation to transactions involving disposal of real property. Where taxpayers had previously reported CGT in the year the contract was executed, they were concerned that the ATO was not properly investigating these before contacting them to assert that CGT had been omitted in the year of property settlement. Other stakeholders noted that the ATO did not seem to consider possible exemptions which may apply to CGT on real property, such as the main residence exemption.

2.3 As a result of this, stakeholders perceived that the ATO's data matching projects selected incorrect cases for review which in turn created additional compliance costs for taxpayers, their agents and the ATO.

2.4 Information provided by the ATO suggests that it does undertake validation checking and verification of both the legislative and non-legislative data and that it also undertakes a process to cleanse and refine such data before they are applied and compared to taxpayer-provided information.64

2.5 Moreover, in 2012-13, the ATO adopted a process whereby all CGT and FSI discrepancy cases selected for audit were manually reviewed by an ATO officer before a discrepancy letter issued to the taxpayer. The purpose of the manual review was to confirm, amongst other things, the existence of the discrepancy which could not otherwise be explained or addressed through the ATO's corporate data holdings and an audit of the taxpayer's return was warranted. The ATO considers that these manual reviews should address some of the concerns outlined above, such as main residence exemptions not being identified or the ATO's system not recognising contract sale dates.65

2.6 While the ATO does not envisage this process to be an ongoing part of its CGT and FSI data matching projects, it considers that learnings from the manual reviews would better inform its business rules and case selection processes to minimise the instance of false positives in these areas.66

Validation checking of data

2.7 The ATO has advised that once data are received, it is subjected to a series of validation checks which are applied to ensure the integrity of the data to be used in matching activities. These checks are based on system rules in the ATO's information technology systems, and are designed to detect obvious errors and missing or duplicate records.67 The range of validation checks applied by the ATO varies depending on the source and nature of the data.68

2.8 While much of the validation checking is automated, the ATO has a number of business rules which identify instances of certain errors occurring frequently throughout a particular data set and removes these for manual review. Other rules exist to require manual review of errors which may not occur frequently but which the ATO considers warrants manual intervention. These errors include where the data set attributes a potentially excessive payment amount to the taxpayer.69

2.9 As the ATO has no control over the format and content of non-legislative data, validation checks are not able to be applied to these data sets. Instead, the ATO undertakes a series of other checks to verify the integrity of the data and to identify missing or incomplete fields and suspicious amounts.

2.10 An example of non-legislative data to which the ATO is unable to apply validation checks is information received from AUSTRAC or foreign revenue authorities for FSI data matching. The reason noted for this is that the ATO is unable to specify the format in which data is provided and as such validation checking of these data sets are not possible.70

2.11 In respect of AUSTRAC information, the IGT observes that while AUSTRAC was established, in large part, to assist the ATO in 'facilitating the administration or enforcement of a taxation law,'71 it is also required to gather information to assist other federal government agencies. These agencies include the Australian Federal Police, the Australian Crime Commission, Customs and the Integrity Commissioner.72

2.12 Given the range of agencies and functions which AUSTRAC information is required to service, the ATO is unable to unilaterally direct AUSTRAC as to the nature and format of information it gathers.

2.13 As a result of the comparatively lower quality and integrity of data matching in non-legislative cases, the limitations on validation checks creates a higher risk of taxpayers being incorrectly identified as having omitted income. In such cases, the ATO needs to adopt an approach which seeks to understand the reasons for discrepancies rather than make any assumptions of non-compliance. Concerns regarding the ATO's approach generally are discussed in Chapter 3.

Identity matching

2.14 Following validation of the data sets received by the ATO, the data is subjected to an identity matching system through the ATO's Client Identification Compliance (CIDC) system. The latter is an automated system which matches the details obtained from third parties against taxpayer details.73

2.15 The primary identity matching indicators used by the ATO are the individual taxpayer's name, address lines, postcode, date of birth, TFN or ABN. The ATO notes that its identity matching systems look to identify exact matches as well as phonetic matches to account for issues such as variant spellings or minor data errors.74 A confidence level is attached to each identity match. These may be high, medium or low. The ATO will also attach an 'unmatched' indicator where the data cannot be matched to any particular taxpayer.75

2.16 The ATO has advised that only data which can be matched to taxpayers with a high level of confidence are progressed through for automated comparison against information reported by taxpayers without further verification.76 The ATO also makes available appropriately identity-matched data for the purposes of pre-filling individual taxpayers' tax returns. The ATO notes that data which return a medium or low level confidence match, or are unmatched, may also be utilised in compliance activities where they have been manually reviewed and the quality of the match can be improved.77

2.17 Identity matching can sometimes be problematic. The ATO has provided real property disposal data as an illustrative example of the importance of, and difficulties associated with, identity matching certain non-legislative data. While it is not the sole cause of low levels of data utility, the ATO notes that some state-based property disposal data lack sufficient identifiers, rendering only about 5 per cent of the data collected useful for its real property CGT data matching. The ATO contrasts this with Queensland's real property disposal data which contains vendor and purchaser date of birth which enable the ATO to more easily identify the parties to the transaction and, accordingly, to use about 80 per cent of the data obtained.78

2.18 The remainder of this chapter discusses the ATO's data matching timeframes and the accuracy of data matching in identifying omitted income.

Timeframes for ATO data matching processes

2.19 Submissions to the IGT expressed concern that delays in the commencement of data matching and contact with taxpayers where discrepancy exists will expose taxpayers to high costs either by way of general interest charge attaching to any amendments or penalties in certain cases. Moreover, such timing issues may also affect the taxpayer's ability to readily access information which may assist to substantiate claims made or address concerns raised by the ATO.

2.20 The IGT notes that by its nature, some delay in data matching is inevitable as it is largely dependent on both information being received from third parties as well as the taxpayers (or their agents) lodging income tax returns.

2.21 ATO information provided to the IGT suggests that, in general, discrepancy letters in relation to data matching for one financial year commence issuing in July of the following financial year. For example, as illustrated in Figure 2, letters in relation to identified 2010-11 discrepancies commenced issuing in July 2012 with exact case numbers being subject of an annual rollout plan that takes into account such factors as resourcing, minimum monetary thresholds and timing of periods of review.79 The latter refers to the statutory period under which the ATO may legally amend an assessment where no fraud or evasion is present.80

Figure 2: Legislative data flowchart

Image of a legislative data flowchart.

View image enlarged

Source: adapted from ATO information

2.22 In relation to non-legislative data, the timeframes are not as established as those outlined above. The main reason for this is that there are no legislated deadlines for the provision of non-legislative data and as such, timeframes are largely dependent on a number of other factors, including its delivery date, destruction rules (that is, how long the ATO is able to keep the data), quality of the data and any work required to refine it. In addition, the availability of resources within the ATO's data warehouse to maintain the collected data needs to be managed.

2.23 Figure 3, below, illustrates the general process adopted by the ATO in relation to non-legislative data matching.

Figure 3: Non-legislative data flowchart

Image of a Non-legislative data flowchart.

View image enlarged

Source: adapted from ATO information

2.24 To illustrate some of the differences in time frames for non-legislative data, the ATO has provided general details in relation to CGT and FSI data timeframes.

CGT data matching timeframes

2.25 The ATO receives CGT real property and share data from state revenue and land title offices, the Australian Securities Exchange and other share registries twice yearly with tranches received in August and February.81

2.26 The ATO notes that each new tranche of information is made available for data matching approximately three months after receipt, and the data matching itself is generally continuous as new data is rolled out and used on an ongoing basis to match against prior year lodgements.82

FSI data matching timeframes

2.27 As noted earlier, the ATO's data collection for FSI data matching purposes originates from a number of sources. Each data source has its own limitations which impact on the focus and timeframes under which FSI data matching operates. For example, the ATO's data warehouse is only capable of storing up to three years' worth of AUSTRAC data and as such, the ATO currently maintains 2010, 2011 and 2012 data and is applying these to its matching projects.83

2.28 FSI data is received from foreign tax authorities throughout the year and frequently relate to transactions which occurred two or three years prior. In addition, AUSTRAC provides data to the ATO on an annual basis. Following appropriate identity matching, this data is made available for the matching process at the request of the ATO's FSI case selection team. These requests are made in accordance with the relevant FSI business rules and requirements and regard is had to the ATO's case actioning resources (that is, matching and selecting a manageable volume of cases).84

IGT observations

2.29 The IGT considers that, ideally, the ATO should undertake legislative data matching within a year following lodgement of the original return and non-legislative data matching on three month cycles as data becomes available. However, there is some time lag between the ATO receiving data and when data matching processes are commenced. Such time lags are as a result of a number of extraneous factors, including legislative due dates for the lodgement of data, the ATO's negotiations with third party data providers to obtain data sets and the ATO's work to cleanse and refine data before any matching is commenced. The latter is particularly pertinent in relation to non-legislative data.

2.30 The IGT also notes that as the ATO maintains up to three years of past non-legislative data within its Data Warehouse, it is possible for the ATO to apply its data matching to verify past years' lodgements, possibly as far back as 2008-09. Moreover, as legislative data forms part of the ATO's corporate data holdings, it is possible for the ATO to compare these to taxpayer reported information going back even further. For example, if a taxpayer lodges their 2005-06 income tax return in 2012-13, it is possible for the ATO to compare the reported information with data from those years in appropriate cases.

2.31 In addition to the impact for taxpayers of having to obtain aged information, delays in the commencement of data matching may also impact on the ATO's compliance work in determining discrepancies and recovering any associated revenue.

2.32 While there may be benefits for taxpayers and the ATO to commence data matching on a more real time basis, this must be balanced against the need for the ATO to ensure, insofar as possible, the accuracy of the data on which it relies. One way in which the ATO could ensure more timely data matching would be to minimise the time taken to validate the data.

2.33 The ATO acknowledges the benefits of commencing data matching activities in a more timely manner but also notes that a number of factors are not within its control, such as legislative timeframes for data reporting or the quality and completeness of data which dictates the amount of work taken to cleanse and refine the data. Therefore, while the ATO accepts that in most cases, data matching may occur within twelve months following lodgement, it is not possible across all data matching projects and that improvements in timeframes may vary as between different data matching projects.

2.34 The IGT notes that the 2013-14 Budget measures should assist to ensure improved data gathering and reporting systems with unique identifiers which would assist in expediting the validation and identity matching processes for a number of reporting regimes. The IGT considers that the ATO could also take more steps to minimise the need for validation by, for example, working closely with third party data providers and focusing its information requests for smaller pools of information with higher levels of accuracy. The latter may also assist to limit the cost exposure for third party data providers.

Recommendation 2.1

The IGT recommends that the ATO minimise the time between data being received and when it may be compared to taxpayer information by regularly reviewing and improving its data validation and identity matching processes.

ATO Response

Agree.

Interest and penalties

2.35 As outlined above, one of the main concerns in relation to delays in data matching is where interest, namely shortfall interest charge (SIC), is imposed and backdated to the original lodgement dates. Where data matching occurs in a timely manner and in line with the ATO's expected timeframes, as discussed above, the imposition of SIC is more manageable. However, where the ATO applies its data matching to prior years, any interest on discrepancies imposed from the original due date for lodgement may create a more significant burden for affected taxpayers.

2.36 Where SIC is imposed, the ATO may choose to remit these in full prior to the finalisation of audit in appropriate circumstances.85 Alternatively, where this does not occur, taxpayers are entitled to apply to the ATO for remission of those charges. The ATO's publicly-stated policy concerning the remission of SIC is outlined in the ATO's Law Administration Practice Statement PSLA 2006/8 Remission of shortfall interest charge and general interest charge for shortfall periods (PSLA 2006/8). This PSLA recognises that in some instances, delay on the part of the ATO may affect the level of interest accrued against a taxpayer and, as such, is a relevant consideration in relation to possible remission.86 Specifically, paragraphs 49 to 51 of that PSLA state:87

Delay in commencing audit

49. The allocation of the ATO's resources necessarily means that not all audits can start immediately following self-assessment by a taxpayer.

50. As a rule of thumb it could be expected that an audit will commence within a period equivalent to 50% of the relevant period of review. However cases may commence at a later time, for example, where information is provided to the ATO from an external source at a later time. Where there has been an unreasonable delay in the ATO allocating a case for audit, remission of interest charges to the base rate may be appropriate for the period of such delay.

51. Where there is an unlimited period to amend an assessment it would normally be appropriate when considering remission to adopt the standard period of review applicable to the year being adjusted.

2.37 The PSLA further outlines a number of different circumstances in which delay on the part of the ATO or the taxpayer may lead to full or partial remissions of SIC88 and those circumstances in which it would not.89

2.38 While it is not immediately apparent on the face of PSLA 2006/8, internal instructions to MEI staff considering SIC remissions indicate that unreasonable delays in relation to data matching may be considered as factors for remission. Specifically, notes accompanying the instructions state:90

Tax officers should also consider remission when circumstances justifying remission are readily apparent (these may not be readily apparent in automatic amendment and data matching cases), or where there may be unreasonable delays in issuing amended assessments.

[Tax officers] should consider whether there were unreasonable delays in all cases where it is readily apparent. Do not broaden the scope of the request to matters which the taxpayer has not contended.

2.39 However, it is unclear in the data matching context what the ATO considers to be an unreasonable delay and whether such delay is 'readily apparent'. Moreover, the internal instructions do not explicitly direct ATO officers to consider whether there was an unreasonable delay in the commencement of data matching activities which may warrant SIC remission, in line with PSLA 2006/8.91

2.40 In respect of the imposition of penalties, the IGT notes that, by and large, the ATO adopts a position of not imposing penalties on data matching cases, as outlined in its discrepancy letters (for bank interest, CGT, FSI and other data matching projects more generally). These letters each contained a variation of the advice that:92

We will not be applying a penalty

If we need to adjust your tax return, we will not apply a penalty. However, if all of your interest income is not included in any future tax return, a penalty will be considered.

2.41 As with SIC, where taxpayers are levied with penalties as a result of data matching activities, they may seek to exercise their right to apply for remission. The ATO has published two practice statements in relation to the remission of penalties. Specifically, these are Law Administration Practice Statement PSLA 2012/4 Administration of penalties for making false or misleading statements that do not result in shortfall amounts and Law Administration Practice Statement PSLA 2012/5 Administration of penalties for making false or misleading statements that result in shortfall amounts. As noted above, the imposition of penalties in data matching cases should be rare, particularly those penalties which turn on culpability.

2.42 However, penalties may apply where data matching identifies income which ought to have been reported but no tax return has been lodged. For example, where the taxpayer has disposed of real property which may result in CGT being reportable but has not lodged their tax return for that financial year, the ATO may issue a lodgement reminder which raises the possibility of penalties and interest being imposed. An example of such a letter states:93

Where we make a default asessment, you may be required to pay tax and may also be liable for:

  • a minimum penalty of 75 per cent of any tax raised from the assessment
  • a penalty for failing to lodge your form on time, and
  • interest charges on any tax you owe.

IGT observations

2.43 The imposition of penalties and interest operates to serve a number of different purposes. In the main, these are designed to deter non-compliant behaviour and encourage compliance with tax lodgement and payment rules. Moreover, the imposition of interest seeks to ensure that taxpayers who do not pay their tax obligations on time do not get an unfair advantage over taxpayers who do and that the revenue is compensated appropriately for any late payments.

2.44 The IGT acknowledges the ATO's advice that in data matching cases, it generally does not impose culpability penalties. However, in respect of SIC which may automatically accrue on adjustments and backdated to original due dates, the IGT is concerned that the level of interest accrued may be unnecessarily punitive where data matching activities involve returns lodged up to three previous years.

2.45 While taxpayers have a right to seek remission of SIC, such applications have associated cost and timing implications. This may be exacerbated where factors the ATO considers relevant to the remission of SIC are unclear, such as delays in the commencement of data matching activities. As outlined earlier, the ATO's binding instruction on SIC remission, namely PSLA 2006/8, when read together with the ATO's internal SIC remission guidelines appears to consider delay in the commencement of audits as a relevant factor.

2.46 However, as the internal guidelines are not public documents, taxpayers or their advisers reviewing PSLA 2006/8 in isolation may not necessarily appreciate that delay in the commencement of data matching may be raised as a relevant factor for the ATO's consideration.

2.47 Accordingly, the IGT considers that there would be benefit in the ATO confirming that data matching audits are analogous to other active compliance activities and update PSLA 2006/8 to clearly state that delay in the commencement of data matching activities is a relevant factor when considering the remission of SIC. To further reinforce this in practice, the ATO should also clarify this position in its internal SIC remission guidelines.

Recommendation 2.2

The IGT recommends that the ATO:

  1. update PSLA 2006/8 to reflect that, consistent with other active compliance activities, delay in commencement of data matching activities is a relevant factor when considering remission of SIC; and
  2. clarify its internal SIC remission guidelines to require ATO officers to consider any delay in the commencement of data matching activities and whether remission is warranted on such grounds.

ATO response

Agree with a qualification

The ATO agrees with the recommendation however is flagging that given the high volume of transactions undertaken through an automated process, consideration of the remission of the Shortfall Interest Charge (SIC) will occur as requested by the taxpayer post issue of the amended assessment.

Accuracy of data matching in detecting omitted income

2.48 After the data has been subjected to both validation and identity checking, an initial comparison is made to determine the population of taxpayer returns with potential discrepancies. This original population may be significant for both legislative and non-legislative data matching as it includes all cases in which the data matching identifies a discrepancy of $1 or more.

2.49 In respect of legislative data matching, the data in Table 6 indicates that in 2011-12, the ATO identified 413,705 and 1,069,855 potential discrepancies relating to salary/wage and interest/dividends respectively. Similarly, non-legislative data matching projects can also involve significant numbers of potential discrepancies. As outlined in Table 7 and Table 8, the ATO identified 1,411,028 potential discrepancies in CGT from the sale of real property in 2011-12 and 1,957,447 in FSI from AUSTRAC for the 2010-11 and 2011-12 financial years.

2.50 Following the initial matching, the ATO compares the above discrepancies against its corporate data holdings to identify any information the ATO already holds that may assist to exclude verified cases from further checking. In addition to this, the ATO also employs a number of business rules to further exclude cases which are unsuitable for audit, unable to the be the subject of audit or uneconomical to pursue. In respect of the latter category, the ATO may issue an advisory letter to taxpayers to inform them that a discrepancy has been identified and to exercise caution in lodging their tax returns for future years.94

2.51 Each data matching project has its own selection and exclusion rules based on the risk sought to be addressed by the ATO.95 By way of illustration, the ATO's current real property CGT business rules exclude for audit those cases in which:96

  • the statutory period for review has expired;
  • the potential revenue gain is below a certain threshold amount;
  • the taxpayer is entitled to a low income tax offset or senior Australian tax offset rendering the potential revenue gain below a certain threshold amount;
  • the main residence exemption applies;
  • the taxpayer is insolvent; or
  • the taxpayer is the subject of current prosecutorial action, their return has been suspended or cancelled or the address of the property is unable to be located.

2.52 Where the results of the application of the above rules are uncertain, manual intervention by ATO officers is required to determine whether a matter should be progressed to audit.97

2.53 In selecting cases for audit, the ATO also has regard to its own resourcing in terms of staffing (including the capability of staff at particular levels) and the numbers of cases which it considers can be completed in a timely manner. Moreover, the ATO also considers whether discrepancies may be addressed through other information available on the ATO's systems as this has a direct bearing on the numbers of matters needing to be manually reviewed.98

2.54 To illustrate the differences between initially identified potential discrepancies and the numbers of cases ultimately selected for audit, the ATO has provided a sample comprising legislative and non-legislative data matching projects (Table 5).

Table 5: Sample of discrepant case pools and finalised cases
Legislative data matching -Investment (interest and dividends) Non legislative data matching - Capital gains tax from the sale of property (individuals) Non legislative data matching - Foreign Source Income from Double Tax Agreement data)
Case actioning year 2012 2012 2012
Tax return year 2011 2008 2007, 2008 & 2009
Number of returns identified with potential discrepancies99
Discrepancies are based on raw data matching processes.
1,069,855 492,093 74,904
Discrepant case pool identified from raw data matching processes100
After the application of business rules
168,322 1,412 5,236
Discrepant returns selected for data matching audits101
After scoping, sampling and data integrity checks
N/A 1,301 3,608
Finalised cases 167,655 1,301 3,608
Active cases 667 0 0
Percentage of case pool finalised 99.60% 100% 100%

Source: ATO

2.55 In addition to the sample above, the ATO has also provided statistics across a range of other data matching activities to show the numbers of potential discrepancies from raw data matches and the number of audits actually completed in certain years. These provide for a more granular examination of the rates of completed audits across a broad sample of projects and are outlined in Tables 6, 7 and 8 below. The ATO has advised that as case selection is refreshed throughout the year, cases completed in a particular year may include those selected from the previous year.

2.56 As shown in Table 6, in approximately half of the projects sampled, the proportion of data matching cases completed is less than 10 per cent of the total potential discrepancies identified. It should be noted that the total potential discrepancies includes very small amounts that would not justify an audit.

2.57 The data matching projects in relation to 'salary and wage', 'Australian Government pension and allowances' and 'employee share schemes' resulted in 6 per cent of total discrepancies being audited. Of greatest concern is the data matching in relation to 'Australian annuities and superannuation income streams' which reported that only about 0.19 per cent of total potential discrepancies were audited.

2.58 In respect of the remaining data matching projects, there were higher rates of cases being completed, with 11 per cent of 'lump sum payments' and 16 per cent of 'investment income' cases being completed. The highest rates of case completion are attributable to the data matching projects in relation to 'Allowances' and 'Australian Government Allowances and payments', reporting 23 per cent and 26 per cent of cases being completed, respectively.

Table 6: ATO legislative data matching - quantity of cases actioned
Subject of data matching Case actioning years Number of returns identified with potential discrepancies Number of data matching cases completed Percentage of cases completed
Legislative Data Matching

Salary and wage

PAYG data is matched to the salary and wage income label on the individual income tax return to check that the taxpayer has returned all of their employment income and claimed the correct tax withheld amounts.

2012 413,705 23,885 6%
2011 459,982 17,519 4%
2010 653,700 13,385 2 %

Lump sum payments

PAYG data is matched to the lump sum payments label on the individual income tax return to check that the taxpayer has returned all of their lump sum payment income and claimed the correct tax withheld amounts.

2012 14,292 1,546 11%
2011 22,485 2,211 10%
2010 24,315 1,908 8%

Allowances

PAYG data is matched to the allowance income label on the individual income tax return to check that the taxpayer has returned all of their allowance income and claimed the correct tax withheld amounts.

2012 152,977 35,679 23%
2011 302,926 40,862 13%
2010 373,855 24,003 6%

Australian Government Allowances and payments

Welfare data is matched to taxpayer returns to check that the taxpayer has returned all of their Australian Government Allowances and payments income.

2012 129,845 34,079 26%
2011 189,413 28,920 15%
2010 176,222 33,189 19%

Australian Government pension and allowances

Welfare data is matched to taxpayer returns to check that the taxpayer has returned all of their Australian Government pension and allowances income.

2012 185,273 11,137 6%
2011 211,885 1,040 0%
2010 191,327 7,125 4%

Australian annuities and superannuation income streams

PAYG data is matched to the Australian annuities and superannuation income streams label on the individual income tax return to check that the taxpayer has returned all of their annuities and superannuation income and claimed the correct tax withheld amounts.

2012 32,686 64 0.19%
2011 65,852 434 1%
2010 88,050 652 1%

Investment income

Interest and dividend data is matched to the gross interest and dividend income labels on the individual income tax return to check that the taxpayer has returned all of their investment income.

2012 1,069,855 167,655 16%
2011 814,579 190,944 23%
2010 804,808 121,954 15%

Employee share schemes Div83A

Employee share scheme data is matched to the employee share scheme income labels on the individual income tax return to check that the taxpayer has returned all of their employee share scheme income and claimed the correct tax withheld amounts.

2012 154,905 8,607 6%
2011 N/A N/A N/A
2010 N/A N/A N/A

Source: ATO

2.59 The statistics in Tables 7 and 8 outline potential discrepancies and case completion rates for CGT and FSI. As in the case of legislative data, the ATO has also advised that the number of cases completed are not analogous to the number of cases selected for audit.

2.60 Tables 7 and 8 show that the population of cases which were completed generally represents an even smaller fraction of total discrepancies when compared with those for legislative data matching. In respect of CGT, all but one of the data matching projects reported case completion rates of less than 1 per cent with only real property CGT for micro enterprises reporting that 7 per cent of potential discrepancies were completed by way of audit.

2.61 Similarly, in respect of FSI, two of the five sampled projects (FSI AUSTRAC and FSI - Non Lodgers) reported that less than 1 per cent of potential discrepancies were actioned with the remaining three projects reporting case completion rates of less than 10 per cent.

2.62 It should be noted that in relation to the OVDI data matching project, the ATO lists a 100 per cent case actioning pool because it made contact by way of initial correspondence with all 9,582 taxpayers within that population. However, the actual proportion of cases completed following this initial contact was 82.5 per cent. The ATO has advised that as this was a specific project to follow up on correspondence that had previously been issued, ATO officers had been directed to specifically action all cases within this pool.

Table 7: ATO non-legislative data matching, CGT - quantity of cases actioned
Subject of data matching Case actioning years Number of returns identified with potential discrepancies Number of data matching cases completed Percentage of cases completed
Non-Legislative Data Matching, CGT

Capital gains tax - from the sale of property

Property disposal data from state and territory revenue offices is matched to taxpayers to identify where a taxpayer has disposed of a property which does not appear to be their main residence and has not returned the capital gains.

2012 1,411,028 2,860 0.20%
2011 Pilot only 418 N/A
2010 Pilot only 158 N/A

Capital gains tax - from the sale of property - non lodger

Property disposal data from state and territory revenue offices is matched to taxpayers to identify where a taxpayer has disposed of a property which does not appear to be their main residence and has not lodged a tax return where it appears that they have an obligation to do so.

2012 280,411 544 0.19%

Capital gains tax - from the sale of property Micro taxpayers

Property disposal data from state and territory revenue offices is matched to taxpayers in the micro market to identify where a taxpayer has disposed of a property which and has not returned the capital gains.

2012 354,102 3,686 7%

Capital gains tax - from the sale of property Micro taxpayers - non lodgers

Property disposal data from state and territory revenue offices is matched to taxpayers in the micro market to identify where a taxpayer has disposed of a property and has not lodged a tax return where it appears that they have an obligation to do so.

2012 55,636 174 0.31%

Capital gains tax - from the sale of shares

Share transaction data provided by share registries is augmented with data from the Australian Security Exchange to identify potential capital gain events from the disposal of shares. This income is then matched to the capital gains tax labels to check that the taxpayer has returned all of their capital gains income from the disposal of shares.

2012 Pilot conducted on single buy single sell 373 N/A

Source: ATO

Table 8: ATO non-legislative data matching, FSI - quantity of cases actioned
Subject of data matching Case actioning years Number of returns identified with potential discrepancies Number of data matching cases completed Percentage of cases completed
Non-Legislative Data Matching, FSI

Foreign source income - from Double Tax Agreement data

Automated exchange data received from treaty parties is matched to the foreign source income label on the individual income tax return to check that the taxpayer has returned all of their foreign source income and claimed the correct tax offset.

2012 74,904 3,608 5%
2011 72,656 5463 8%

Foreign source income - AUSTRAC

Incoming amounts listed in AUSTRAC are matched against foreign source income label on the individual income tax return to check that the taxpayer has returned all of their foreign source income.

2012 1,957,447 1,746 0.09%
2011 3.875 0.19%

Foreign source income - Mislabelling

Individual income tax returns are identified where a taxpayer has returned an amount of foreign source income at a non-assessable label on and no amount at the assessable label. These returns are checked to ensure that the taxpayer has returned all of their foreign source income

2012 156,433 5,164 3%

Foreign source income - Offshore Voluntary Disclosure Initiative (OVDI)

Follow up work in relation to non-responses to the Offshore Disclosure project. The Offshore Disclosure project identified that these received incoming amounts via AUSTRAC. Where no response was received to the initial voluntary disclosure letter a data matching letter was issued.

2012 9,582 7,909 100% of taxpayers were contacted under this project 82.5% of cases actioned following this initial contact

Foreign source income - non lodgers

Automated exchange data received from treaty parties and incoming amounts listed in AUSTRAC are used to identify resident taxpayers that appear to have a requirement to lodge a return due to the amount of foreign source income they have received.

2012 16,927 15 0.09%

Source: ATO

IGT observations

2.63 The IGT acknowledges that the number of cases completed in a particular year do not equate to those selected for audit. However, the IGT also notes from the sample in Table 5, that the levels of cases selected for audit represent only a small fraction of the potential discrepancies from raw data matches. The low level of cases actually selected for audit may be reflective of a number of different factors.

2.64 Firstly, through the use of its business rules, the ATO is able to reject cases which do not yield a revenue gain, are below pre-set thresholds for particular projects, do not pose a sufficiently high level of risk or do not otherwise warrant an audit.

2.65 Secondly, it may suggest that tax returns initially identified as containing discrepancies are later found to be not so. This may be due to discrepancies being explained or verified through data matching with other information held by the ATO. In the case of CGT and FSI more recently, it may also be due to the ATO's more rigorous manual reviews of these discrepancies before selecting cases for audit.

2.66 Thirdly, it may indicate that the third party data obtained and relied upon by the ATO only identifies instances of discrepancies which are ultimately not suitable for audit. This suggests some scope for the ATO to better refine its data requests to obtain more focused information to assist in its data matching.

2.67 Fourthly, where business rules and other review processes reduce the numbers of cases suitable for audit, it may suggest that the risk areas on which the ATO is focusing its efforts and resources need to be reconsidered. Where, for example, in the case of 'Australian annuities and superannuation income streams', the case completion rate over three years has not exceeded 1 per cent, it is important to question whether ongoing examination and commitment of resources is warranted.

2.68 The IGT notes that risks may change over time and the ATO needs to be responsive in its evaluation of risk areas warranting examination. Such an examination should consider the nature of the risk being investigated, the data on which the ATO relies and strategies which the ATO could employ to request and make use of more and better data to address the risk in question.

2.69 Through discussions with the IGT, the ATO has indicated that it is mindful of the need to ensure that it's data matching projects are effective and generate a return on investment. It notes that with the benefit of experience over many years in relation to its data matching program, its case selection is designed to canvass a broad mix of projects and risk areas which it expects will yield the highest return on its staff and resource commitments.

2.70 The ATO's rollout plans for data matching case selection record both expected and actual number of cases selected and completed each week. Through these plans, the ATO is able to assess its return on investment through measuring staff costs and other incidental expenses against revenue expected to be raised.102

2.71 The progress of data matching case completion is monitored by senior directors within the DMCS on a weekly basis. These directors assess expected and actual case selection and completion for particular projects. This assessment involves consideration of such factors as whether certain projects are progressing more efficiently, whether thresholds may be lowered to action more cases, resources deployed to other projects which are not performing to expected timeframes or commencing audits relating to the following year's data matching cases.103

2.72 Whilst the weekly discussions amongst the senior directors provide a timely view of the ATO's progress, it does not appear that the ATO consolidates this intelligence to examine and analyse longer term trends and impacts.

2.73 The IGT believes that there is room for improvement in assessing the longer term effectiveness of data matching projects. For example, during the course of this review, the IGT has not seen formalised assessment processes which may determine, amongst other things, whether a particular data matching project should continue if the levels of cases being identified for audit are low and the levels of revenue being protected are disproportionate to the cost of the project.

2.74 The IGT considers that it is important for the ATO to undertake formal evaluations of this kind, particularly where certain projects are not performing as expected or appear to be declining in their effectiveness. Such evaluations would, for example, enable the ATO to redirect its resources to projects with better yields or seek further liaison with third party providers or legislative change to obtain better data.

Recommendation 2.3

The IGT recommends that the ATO:

  1. formalise and improve its current processes for determining the reasons for compliance action being only undertaken in a very small number of cases where potential discrepancies have been identified between third party information and those contained in corresponding tax returns;
  2. following the above investigation, seek to improve the effectiveness of the relevant data matching projects through, for example, improving the quality and utility of third party data; and
  3. where (b) is not possible, consider abandoning the relevant data matching projects and redirecting those resources to other data matching projects that are likely to yield better results.

ATO response

Agree

The ATO agrees with this recommendation. Awareness of, and adaption to new opportunities will help keep our data matching efforts relevant and effective.

Strike rates of ATO data matching activities

2.75 A key measure of the success of the ATO's data matching activities is the 'strike rate'. That is, the number of cases selected for audit, in which the ATO identifies instances of omitted or underreported income and makes an adjustment, as measured against the population of cases selected for audit.104

2.76 A sample of strike rates, and corresponding false positive rates,105 for legislative data matching activities is presented in Table 9, below. It should be noted that this data does not take into account any adjustments later reversed on review. Rates of adjustments and reversals are discussed in Chapter 4.

2.77 As indicated in the table below, the ATO's legislative data matching, with a few exceptions, results in very high strike rates. This is most evident in relation to the 'Allowances' and 'Australian Government Allowances and Payments' data matching projects which, in 2011-12, reported strike rates of 98 per cent and 99 per cent respectively.

2.78 Of the remaining data matching projects, 'Investment Income' and 'Employee Share Scheme' also experienced very high strike rates of 94 per cent and 95 per cent respectively in 2011-12. A lower, though still considerable, strike rate was reported in relation to the ATO's 'Salary and wage' and 'Lump sum payments' projects.

2.79 While the sample of legislative data matching activities generally reported high strike rates, this was not a consistent outcome. In relation to those projects which data matched 'Australian Government pension and allowances' and 'Australian annuities and superannuation income streams', the strike rates were considerably lower than those discussed above (68 per cent and 31 per cent, respectively).

2.80 Moreover, an examination of these strike rates over a three year period shows significant decline in the strike rates. Specifically, the strike rate for 'Australian Annuities and superannuation income streams' fell from 94 per cent in 2009-10 to 73 per cent in 201011 to the 31 per cent discussed above. The 'Australian government pension and allowance' project experienced a small reduction in the strike rate from 2009-10 (96 per cent) to 2010-11 (93 per cent) followed by a decrease to the previously mentioned 68 per cent.

2.81 The ATO has advised that in relation to the superannuation income streams data matching project, the significant decline in strike rate was attributable to a the lack of manual reviewing of cases before selection for audit after 2009-10.

Table 9: Legislative data matching - strike rates
Subject of data matching Case actioning years Number of data matching cases completed Number of adjustments Strike rate (percentage of cases with adjustment) Number of cases requiring no further action (no adjustment required Rate of no further action
Legislative Data Matching

Salary and wage

PAYG data is matched to the salary and wage income label on the individual income tax return to check that the taxpayer has returned all of their employment income and claimed the correct tax withheld amounts.

2012 23,885 21,347 89% 2,538 11%
2011 17,519 14,871 85% 2,648 15%
2010 13,385 10,519 79% 2,866 21%

Lump sum payments

PAYG data is matched to the lump sum payments label on the individual income tax return to check that the taxpayer has returned all of their lump sum payment income and claimed the correct tax withheld amounts.

2012 1,546 1,253 81% 293 19%
2011 2,211 1,329 60% 882 40%
2010 1,908 1,829 96% 79 4%

Allowances

PAYG data is matched to the allowance income label on the individual income tax return to check that the taxpayer has returned all of their allowance income and claimed the correct tax withheld amounts.

2012 35,679 34,834 98% 845 2%
2011 40,862 39,636 97% 1,226 3%
2010 24,003 22,678 94% 1,325 6%

Australian Government Allowances and payments

Welfare data is matched to taxpayer returns to check that the taxpayer has returned all of their Australian Government Allowances and payments income.

2012 34,079 33,612 99% 467 1%
2011 28,920 28,631 99% 289 1%
2010 33,189 32,696 99% 493 1 %

Australian Government pension and allowances

Welfare data is matched to taxpayer returns to check that the taxpayer has returned all of their Australian Government pension and allowances income.

2012 11,137 7,534 68% 3,603 32%
2011 1,040 965 93% 75 7%
2010 7,125 6,809 96% 316 4%

Australian annuities and superannuation income streams

PAYG data is matched to the Australian annuities and superannuation income streams label on the individual income tax return to check that the taxpayer has returned all of their annuities and superannuation income and claimed the correct tax withheld amounts.

2012 64 20 31% 44 69%
2011 434 317 73% 117 27%
2010 652 616 94% 36 6%

Investment income

Interest and dividend data is matched to the gross interest and dividend income labels on the individual income tax return to check that the taxpayer has returned all of their investment income.

2012 167,655 158,381 94% 9,274 6%
2011 190,944 181,158 95% 9,786 5%
2010 121,954 110,609 91% 11,345 9%

Employee share schemes Div83A

Employee share scheme data is matched to the employee share scheme income labels on the individual income tax return to check that the taxpayer has returned all of their employee share scheme income and claimed the correct tax withheld amounts.

2012 8,607 8,193 95% 414 5%
2011 N/A N/A N/A N/A N/A
2010 N/A N/A N/A N/A N/A

Source: ATO

2.82 The ATO's non-legislative data matching projects generally experience lower strike rates when compared with legislative data matching. As observed in Table 10 which provides a sample of strike rates from different data matching projects concerning CGT, only two projects yielded strike rates which are comparable to those reported for legislative data matching. These are the 'CGT for the sale of real property non-lodgers' and 'CGT from the sale of shares.'

2.83 The remaining CGT data projects all reported strike rates of 60 per cent or lower in 2011-12, with the lowest being in relation to CGT for the sale of real property by micro enterprise taxpayers (31 per cent).

2.84 As the majority of this sample involves CGT data matching projects which only commenced in 2011-12, a three year comparison is not possible. The only exception to this is the data matching project concerning CGT from the sale of real property by individual taxpayers. Over a three year period, the strike rate for this project has dropped considerably from 91 per cent in 2009-10 to 69 per cent in 2010-11 and 52 per cent in 2011-12.

Table 10: Non-legislative data matching, CGT - strike rates
Subject of data matching Case actioning years Number of data matching cases completed Number of adjustments Strike rate (percentage of cases with adjustment) Number of cases requiring no further action (no adjustment required) Rate of no further action (percentage of cases with no adjustment)
Non-Legislative Data Matching, CGT

Capital gains tax - from the sale of property

Property disposal data from state and territory revenue offices is matched to taxpayers to identify where a taxpayer has disposed of a property which does not appear to be their main residence and has not returned the capital gains.

2012 2,860 1,475 52% 1,385 48%
2011 418 288 69% 130 31%
2010 158 143 91% 15 9%

Capital gains tax - from the sale of property - non lodger

Property disposal data from state and territory revenue offices is matched to taxpayers to identify where a taxpayer has disposed of a property which does not appear to be their main residence and has not lodged a tax return where it appears that they have an obligation to do so.

2012 544 508 93% 36 7%

Capital gains tax - from the sale of property Micro taxpayers

Property disposal data from state and territory revenue offices is matched to taxpayers in the micro market to identify where a taxpayer has disposed of a property which and has not returned the capital gains.

2012 3,686 1,159 31% 2,527 69%

Capital gains tax - from the sale of property Micro taxpayers - non lodgers

Property disposal data from state and territory revenue offices is matched to taxpayers in the micro market to identify where a taxpayer has disposed of a property and has not lodged a tax return where it appears that they have an obligation to do so.

2012 174 105 60% 69 40%

Capital gains tax - from the sale of shares

Share transaction data provided by share registries is augmented with data from the Australian Security Exchange to identify potential capital gain events from the disposal of shares. This income is then matched to the capital gains tax labels to check that the taxpayer has returned all of their capital gains income from the disposal of shares.

2012 373 368 99% 5 1%

Source: ATO

2.85 Similar to the CGT data matching projects, those concerning FSI also report less consistent strike rates when compared with legislative data matching activities. These strike rates are outlined in Table 11 below.

2.86 The highest strike rates were reported in relation to the non-lodgement project which produced a 100 per cent strike rate from a case pool of 15. The data matching of double tax agreement information, AUSTRAC information and mislabelled information yielded 88 per cent, 73 per cent and 54 per cent strike rates respectively.

2.87 The lowest strike rate observed in the same sample related to the ATO's OVDI which took place at the end of 2011 and in early 2012. The project involved ATO officers making contact with taxpayers who had previously received letters from the ATO inviting them to make voluntary disclosures in relation to offshore holdings. This project only yielded a strike rate of 54 per cent.

Table 11: Non-legislative data matching, FSI - strike rates
Subject of data matching Case actioning years Number of data matching cases completed Number of adjustments Strike rate (percentage of cases with adjustment) Number of cases requiring no further action (no adjustment required) Rate of no further action (percentage of cases with no adjustment)
Non-Legislative Data Matching, FSI

Foreign source income -from Double Tax Agreement data

Automated exchange data received from treaty parties is matched to the foreign source income label on the individual income tax return to check that the taxpayer has returned all of their foreign source income and claimed the correct tax offset.

2012 3,608 3,171 88% 437 12%
2011 5463 4,862 89% 601 11%

Foreign source income -AUSTRAC

Incoming amounts listed in AUSTRAC are matched against foreign source income label on the individual income tax return to check that the taxpayer has returned all of their foreign source income.

2012 1,746 1,274 73% 472 27%
2011 3,875 3,333 86% 543 14%

Foreign source income -Mislabelling

Individual income tax returns are identified where a taxpayer has returned an amount of foreign source income at a non-assessable label on and no amount at the assessable label. These returns are checked to ensure that the taxpayer has returned all of their foreign source income

2012 5,164 4,125 80% 1,039 20%

Foreign source income -Offshore Voluntary Disclosure Initiative (OVDI)

Follow up work in relation to non-responses to the Offshore Disclosure project. The Offshore Disclosure project identified that these received incoming amounts via AUSTRAC. Where no response was received to the initial voluntary disclosure letter a data matching letter was issued.

2012 7,909 Letters were issued to all 9,582 taxpayers 4,267 54% 3,642 46%

Foreign source income -non lodgers

Automated exchange data received from treaty parties and incoming amounts listed in AUSTRAC are used to identify resident taxpayers that appear to have a requirement to lodge a return due to the amount of foreign source income they have received.

2012 15 15 100% 0 0%

Source: ATO

2.88 The ATO has acknowledged the declining strike rates experienced in both the CGT and FSI data matching projects, particularly in 2011-12. In the case of FSI, the strike rates continued to decline in 2012-13. However, the ATO has advised that since the introduction of manual reviews for both FSI and CGT audit case selection, it has seen an improvement in strike rates in 2012-13 and 2013-14 year to date. These improved strike rates, together with other global strike rates for the data matching program generally, are outlined in Table 12.

Table 12: Global strike rates for data matching program
2009-10 2010-11 2011-12 2012-13 2013-14 YTD
Legislative 92% 95% 93% 93% 91%
Legislative - Other106 56% 53% 82% 73% 85%
Non-Legislative - CGT 91% 69% 47% 63% 78%
Non-Legislative - FSI N/A 88% 70% 55% 81%
Non-Legislative - Other107 60% 73% 88% 84% 94%
Overall Strike Rate 92% 91% 90% 79% 89%

Source: ATO

IGT observations

2.89 Based on the statistics provided by the ATO in Table 9, Table 10 and Table 11, the IGT notes that, with a few exceptions, the ATO's initial strike rates for both legislative and non-legislative data matching projects are high. The statistics in Table 12 demonstrate that data matching, generally, yields positive outcomes and strike rates. Notwithstanding, the declines prior to 2011-12 in relation to CGT, and 2012-13 for FSI, the strike rates for both appear to have improved in the 2013-14 year to date.

2.90 It is clear from a comparison of the two types of data matching that non-legislative data matching generally results in lower strike rates than legislative data. Moreover, and as discussed earlier, the proportions of cases actually actioned from non-legislative data are a smaller fraction of the total number of potential discrepancies identified. This suggests that non-legislative data obtained by the ATO does not contain all of the information needed by the ATO to address the risks it has identified.

2.91 Where the ATO actions cases with a lower strike, this has the potential to impose additional compliance costs on otherwise compliant taxpayers who are required to engage tax agents or otherwise spend time responding to ATO enquiries. Moreover, where the ATO requests large amounts of information which are ultimately not used, this creates costs for the third party providers and administrative costs for the ATO.

2.92 The ATO may address these issues by examining whether there are strategies to reduce the amount of information it seeks from third parties or improving its data validation strategies to make greater use of the data it receives. However, there may be an inherent difficulty in the ATO seeking to request less but more focused information. The ATO only identifies the risk which it seeks to treat but is unable to specifically identify the taxpayers from whom data should be sought. The role of the Data Matching Steward, discussed above, is to develop a strategic approach to the collection and application of data in active compliance activities, assist the business lines to identify the best sources of data and how these may be requested while minimising the impact on those third parties. There is a risk that, if the ATO sought more specific and targeted information, those third party providers would be required to spend more time and costs in responding to the ATO's request.

2.93 The IGT notes that the 2013-14 Budget Measures should assist the ATO in engaging with certain third parties (such as State land titles offices and AUSTRAC) to develop systems to capture more specific and targeted information. While the extent and details of these measures are not yet public, this move provides the ATO with an ideal opportunity to reassess and improve its current data matching framework.

2.94 As the ATO moves to expand its data matching function, it should take this opportunity to strengthen its organisational capability to enhance its administration of data matching generally. In doing so, the ATO should develop a strategic enterprise approach to the collection and application of third party data in its future data matching activities, including the use of manual reviews for audit selection in relation to new risks or new data sources. In the interim, there is scope for the ATO to engage with existing third party providers to examine ways in which information requests may be better targeted to ensure that more relevant information is obtained.

2.95 In addition to this, and having regard to the decreasing trend both in legislative and non-legislative data matching strike rates, the ATO should continuously review and evaluate the effectiveness of its data matching projects.

2.96 Information provided by the ATO suggests that it does undertake some evaluation and assessment of its data matching programs. These include assessments as part of the ATO's Integrated Quality Framework (IQF), a high level assessment of a sample of completed cases against nine standard criteria,108 as well as formal evaluations required under the AIC's data matching guidelines.109

2.97 The IGT notes, however, that these evaluations do not specifically address issues concerning the strike rates for different data matching projects and the causes of variances in these strike rates. Whilst the ATO does not undertake formal evaluations of its data matching projects, as stated earlier, its senior directors within the DMCS meet on a weekly basis to review and discuss progress in different data matching projects.

2.98 As the data matching program continues to expand, the IGT considers that the ATO needs to identify the areas in which its data matching work is effective and those in which it is less effective. This information would enable the ATO to improve its processes to minimise the occurrences of taxpayers being incorrectly identified through data matching, thereby limiting compliance costs and reducing the resource impacts on the ATO.

Recommendation 2.4

The IGT recommends that the ATO:

  1. formalise a strategic enterprise approach to the collection and application of third party data used in its data matching activities. Such an approach should include:
    1. undertaking pilots and/or manual reviews for audit selection where new data sources are being used or new risks sought to be treated; and
    2. consulting with third party data providers regarding their 'natural business systems' and any necessary changes to these systems or processes to accommodate ATO needs and undertaking a cost-benefit analysis to determine whether the ATO can reimburse or subsidise any of their associated costs.
  2. consolidate its evaluations of data matching projects to capture observations made during those projects to determine:
    1. any material changes in the effectiveness of the project to accurately identify taxpayers who had omitted or underreported their income;
    2. the underlying causes for the variation in the data matching project's effectiveness; and
    3. use this information to enhance similar data matching projects for future years, or to redeploy staff to focus on other data matching projects representing higher risks.

ATO response

Agree

The ATO agrees to formalise our strategic enterprise approach to the collection and application of third party data used in our data matching activities and consolidate our evaluations to capture observations made during those projects.

Accuracy of data matching used in pre-filling

2.99 Data collected by the ATO is not only used in data matching to detect omitted or underreported income. The ATO also applies its corporate data holdings to pre-fill electronic income tax returns where appropriately identity-matched to specific taxpayers. The ATO notes the following benefits of pre-filling for both taxpayers and for itself:110

The benefits for you

  • Pre-filling makes doing your e-tax tax return easier and more accurate.
  • Pre-filling downloads information to partially complete your e-tax return for you, ready for you to review.
  • Most information goes directly to the correct items on your e-tax return, so you don't have to work out where to put it.
  • Other useful information is downloaded into a summary for you to review to help remind you about amounts you might need to declare or claim on your e-tax tax return.
  • Provided you resolve any discrepancies between your records and what is pre-filled, you know the income on your e-tax return matches what we have on our records.
  • You have the convenience of access to the pre-filling service 24 hours a day, seven days a week (during the period e-tax is available).
  • You can identify lost or forgotten records, such as lost bank statements and payment summaries sent to previous addresses.
  • If all your information does not download you can simply add in any missing details or register for the alerts service in e-tax and we will let you know when your information is available.

The benefits for us

  • There will be fewer discrepancies for us to follow up because the information is correctly entered on your e-tax return.
  • There will be fewer amendments for us to process because you can confirm that the information you have and the information from third-party providers is correct.

2.100 Taxpayers and tax agents have welcomed the ATO's use of corporate data to assist taxpayers in the completion of their income tax returns. However, and notwithstanding this support, some tax agents have expressed concern that taxpayers sometimes take the pre-filled information as being complete and accurate without further checking to assure themselves. The reasons for such inaccuracy or incompletion include:

  • general delays in some entities reporting to the ATO or where information has not been received within statutory timeframes for lodgement;
  • the ATO is unable to accurately identity match the information which has been reported to a particular taxpayer;
  • the ATO's identity matching has allocated certain data to a particular taxpayer inaccurately; or
  • the ATO has returned data sets to the information providers owing to errors or other inaccuracies which have been observed.

2.101 Another reason for incomplete or inaccurate pre-filled data is that the ATO does not apply stringent validation to these data sets before making them available for pre-filling.111 In part, this turns on the need for the ATO to make such data available as soon as possible during Tax Time so that taxpayers may make use of it in lodging their income tax returns.

2.102 Moreover, the ATO has indicated that data sets which are made available for pre-filling purposes are usually legislatively based which outline specific reporting requirements. In addition, the ATO considers that the integrity of such data sets is high owing to the many years of interaction between the data providers and the ATO.112

2.103 The ATO has advised that where it has received data which it considers less reliable, it only makes these available to tax agents via the Tax Agent Portal for pre-filling purposes. In addition to this, where the ATO identifies issues with existing pre-filled data sets, it will withdraw these from pre-filling, publish details of the issues on its website113 and request updated data sets from the third party provider.114

IGT observations

2.104 Where taxpayers accept pre-filled data as complete and accurate, they may not actively review their records to ensure correct figures have been pre-filled so as to reduce the likelihood that a discrepancy will arise during post-issue data matching activities.

2.105 However, the ATO website does caveat that 'pre-filling does not alter your current responsibilities to provide a complete and accurate income tax return' and 'all you have to do is check the pre-filled information and add any missing details.'115 These references are slight when compared to the other references suggesting that pre-filling enhances the accuracy of taxpayers' returns which may in turn lead taxpayers to accepting pre-filled data without further checking. Similar warnings are given to taxpayers when using the e-tax software.

2.106 The IGT acknowledges that the taxpayer bears the onus of confirming the veracity of any information reported in their tax return. However, it is important to note that this is not always easily accomplished. The ATO recognises that, on some occasions, pre-filled information may contain errors or may not accurately reflect the taxpayer's circumstances. In such situations, the ATO has outlined on its website a number of different steps the taxpayer may wish to take. These include:116

  • checking the pre-filled information against the taxpayer's own records and statements;
  • referring to the ATO's 'known issues' webpage for information on steps which may be taken in respect of specific issues;117
  • addressing any perceived inaccuracies with the information provider; and
  • amending the pre-filled information before lodging the income tax return.

2.107 Stakeholders have indicated to the IGT some of the difficulties associated with trying to confirm the accuracy of ATO data. In one case, a taxpayer queried the amount of bank interest which had been pre-filled in their income tax return. The pre-fill report provided only three digits of the relevant bank accounts. The taxpayer made enquiries with the bank but was advised that unless more complete account numbers were provided, the bank was unable to assist. When the taxpayer requested more detailed account information from the ATO, the ATO advised that owing to secrecy and privacy requirements, the ATO was unable to provide the requested details. This is so as the taxpayer had advised the ATO the account did not belong to them.

2.108 Such a situation places the taxpayer in a difficult position. The taxpayer is not being able to clarify the reasons why these amounts had been pre-filled on their account and to have the data corrected accordingly. This leads to two undesirable options. The taxpayer may either accept the incorrect pre-filled data, or reject it with the associated risk that a discrepancy may be identified when the ATO commences its data matching.

2.109 The ATO also recognises this risk. Its website cautions taxpayers that:118

Avoid an audit - resolve discrepancies before you lodge

If you amend the pre-filled information on your e-tax return, it is important to resolve any discrepancies you identify with the organisation that provided that information to us before you lodge.

This is because we routinely do information matching to identify discrepancies between the information on a lodged tax return and the information provided to us by external information providers.

When a discrepancy is identified, we will seek clarification from you.

2.110 The ATO's internal instructions to its staff provide an escalation process to manage cases in which the taxpayer claims that the pre-filled data does not belong to them. Specifically in relation to bank interest and ownership of bank accounts, the internal instructions note:119

There have been occasions where the customer claims to have no knowledge of the account information reported. The customer should follow this up with the financial institution that reported the information to us to clarify ownership of the account. If the customer insists that the account does not belong to them, escalate the enquiry.

2.111 The ATO has advised that these matters are escalated to the Third Party Data Management team for investigation. It has also further advised that where taxpayers are able to satisfy proof of identity, the ATO officer is to provide full details of relevant data to the taxpayer so that they may make appropriate enquiries with the information provider. Where proof of identity is not satisfied, the ATO considers that it is precluded from disclosing any further specific information by existing secrecy and privacy laws.

2.112 The IGT considers that where these escalated cases are actioned in accordance with the ATO's internal guidelines, situations such as those outlined above should generally be avoided. However, problems such as the one raised above only arise if processes are not followed and taxpayers are not provided with sufficient information.

2.113 In order to minimise the uncertainty for taxpayers who contact the ATO in relation to pre-filling issues, the ATO should widely communicate its processes for managing these enquiries. Moreover, it is essential that the ATO reinforce these procedures with its staff, across all relevant areas, to ensure that such enquiries are effectively managed and, in appropriate cases, assisting taxpayers to resolve any inaccuracies or incompleteness in pre-filled data in a timely and cost-effective manner.

Recommendation 2.5

The IGT recommends that the ATO:

  1. widely communicate a streamlined process through which taxpayers and tax agents may clarify or correct pre-filled data which is incomplete or inaccurate; and
  2. periodically reinforce instructions and escalation processes for ATO staff to manage these enquiries and, where appropriate, assist taxpayers to resolve them in a timely manner.

ATO response

Agree


64 Above n. 15.

65 ATO, 'Case selection process overview', July 2012, internal ATO document; ATO, 'FSI AUSTRAC case selection process', July 2012, internal ATO document; ATO, 'FSI case selections fully discrepent DTA case pool', July 2012, internal ATO document; ATO, 'Example of manual checks on CGT cases before selection', February 2013, internal ATO document.

66 Above n. 22.

67 Above n. 15.

68 Ibid.

69 ATO, 'AIIR Lodgment Summary and Automatch', internal ATO document, pp. 18 - 20.

70 Above n. 15.

71 Section 125 of the Anti-Money Laundering and Counter-Terrorism Financing Act 2006.

72 Explanatory Memorandum, House of Representatives, Anti-Money Laundering and Counter-Terrorism Financing Bill 2006, pp. 135-136.

73 ATO, 'An Executive Overview of ATO Identity Matching Software; ATO, Data Matching and Integrity, ATO Identity Matching Outcomes', internal ATO document.

74 Ibid.

75 Above n. 15.

76 Ibid.

77 Above n. 22.

78 Above n. 16.

79 ATO, communication with the IGT, 18 June 2013.

80 Section 170 of the Income Tax Assessment Act 1936.

81 Above n. 15.

82 Ibid.

83 Ibid.

84 Above n. 15.

85 Commonwealth Ombudsman, Submission to the Joint Committee of Public Accounts and Audit Annual Hearing with Commissioner of Taxation, September 2012, p. 5.

86 Law Administration Practice Statement PSLA 2006/8 Remission of shortfall interest charge and general interest charge for shortfall periods, paras. 43 - 80.

87 Ibid., paras. 49 - 51.

88 Ibid., paras. 58 - 60, 62.

89 Ibid., paras 64 and 74.

90 ATO, 'Shortfall Interest Charge - Decision Making Tool', internal ATO document, steps D and G, p. 1.

91 Above n. 86, para. 50.

92 ATO, Interest advisory letter, BAU.

93 ATO, CGT Lodgement reminder letter.

94 Above n. 22.

95 Above n. 15.

96 ATO, 'Information Matching System, Business Requirements and Rules, CGT', internal ATO document, pp. 3-4.

97 Above n. 96.

98 Above n. 16.

99 The ATO has advised that this raw data matching process includes label to label discrepancies where the discrepancy value is equal to or greater than $1. At this stage, business rules have not been applied.

100 The ATO has advised that at this stage, the impact of rebates, offsets, potential tax free thresholds and other rules have been applied.

101 This refers to the ATO's manual review processes of CGT and FSI cases before selecting cases for audit.

102 Above n. 22.

103 Ibid.

104 Manish Gupta and Vishnuprasad Nagadevara, 'Audit selection strategy for improving tax compliance -application of data mining techniques' in Ashok Agarwal and V. Venkata Ramana (eds), Foundations of E-government (2007) p. 383.

105 Ibid. A false positive occurs where data matching identifies a particular case as potentially having omitted income which is later shown to not be true.

106 These projects include those relating to Medicare Levy exemption, Employee Share Schemes and Partnerships, Trust Distributions and Unit Trust Distributions.

107 These relate to non-legislative projects other than CGT and FSI, such as dependent spouse offset and education tax refund.

108 See for example, ATO, 'Micro Enterprises and Individuals Data Matching and Compliance Integrated Quality Framework May-June Report' August 2012, internal ATO document.

109 See for example, ATO, 'Data Matching Program Evaluation' 19 April 2012, internal ATO document.

110 ATO, Pre-filling services - etax, 12 August 2013.

111 ATO, Communication with the IGT, 22 July 2013.

112 Ibid.

113 ATO, Pre-filling known issues, 19 July 2013.

114 Above n. 111.

115 Above n. 110.

116 ATO, What if pre-filled information is wrong, 12 August 2013.

117 Above n. 113.

118 Above n. 116.

119 ATO information provided on 27 June 2013.