Show download pdf controls
  • Methodology

    The large corporate groups income tax gap is derived through applying a model-based bottom-up approach.

    There are five steps in applying the model-based bottom-up methodology to estimate the large corporate groups income tax gap. These steps are expanded upon below followed by a summary of the overall estimate:

    Find out about:

    Step 1: Calculate amendments

    The results of amendments, both ATO and client-initiated, are used to estimate the tax gap for the entire population. We use:

    • the actual result of compliance activities, including the amendments from completed audits and reviews
    • taxpayer voluntary disclosures
    • expected future compliance outcomes for material amounts in dispute
    • projected future amendments.

    We project future amendments to account for the time delay between a tax return being lodged, and any final amendments that will be made. As complex tax cases may take a number of years to resolve, the amendments may not be received until several years after the tax return was lodged.

    To account for these future amendments, we use data on the value and timing of past amendments to project amendments we are likely to receive in the future. As we revise the gap in future years, we will use refreshed amendment information to update our amendment results and to improve future projections.

    We then aggregate the amendments, including projected amendments, for the population to determine the total amendment result.

    Step 2: Integrate tax assured data

    We use our tax assured data in our estimation. This allows us to more accurately calculate unreported tax and derive a figure for non-detection.

    For large corporate groups we assure tax by collecting evidence directly from taxpayers.

    More information about our approach is in Tax assured: gaining confidence the right amount of tax is reported.

    Step 3: Calculate unreported tax

    Unreported tax is the additional tax we estimate may be raised if we were to undertake compliance activities on the entire population of large corporate groups.

    To estimate unreported tax, we calculate adjustment factors based on actual and projected future amendments.

    These factors are then discounted to account for selection bias. This reflects that our compliance activities are biased towards areas of higher risk than the risk level in the general population.

    We then apply these factors to each entity in the population to estimate the total amount of unreported tax. The factors may be discounted where the tax paid by the entity has been assured, reflecting our higher confidence in those amounts of tax paid.

    Step 4: Estimate non-detection

    We uplift the estimates from the preceding steps to account for non-compliance that is not detected through our compliance activities. We do this by applying uplift factors to the tax amounts based on the level of tax assurance.

    Given the confidence we have in tax amounts assured through our justified trust program, we apply a lower non-detection factor to those amounts compared to amounts we have not assured.

    Find out more in Ensuring complete estimates: Non-detection.

    Step 5: Estimate theoretical liability – gross gap and net gap

    We add total amendments (Step 1c), unreported tax (Step 3) and non-detection (Step 4) to determine the gross gap. We then add the amount of tax voluntarily reported and paid in order to calculate the theoretical liability. We then subtract total amendments from the gross gap to determine the net gap.

    Estimate summary

    Table 2 provides a summary of each step of the estimate for each year. It shows the calculation for each of the steps described from 2011–12 to 2017–18. Steps 1 through to Step 5d are in dollar values, and Steps 5e and 5f are in percentage values.

    Table 2: Summary of large corporate groups income tax gap estimation process

    Step

    Description

    2011–12

    2012–13

    2013–14

    2014–15

    2015–16

    2016–17

    2017–18

    1a

    Amendments ($m)

    1,205

    1,659

    1,127

    817

    737

    574

    499

    1b

    Projected amendments ($m)

    192

    334

    627

    1,006

    1,279

    1,449

    1,604

    1c

    Total amendments ($m)

    1,397

    1,993

    1,754

    1,823

    2,016

    2,023

    2,103

    2

    Tax assured ($m)

    77

    743

    4,629

    7,303

    25,979

    26,781

    26,078

    3

    Unreported tax ($m)

    1,198

    1,574

    1,403

    1,063

    885

    791

    915

    4

    Non-detection ($m)

    1,383

    1,293

    1,240

    1,188

    702

    908

    1,119

    5a

    Gross tax gap ($m)

    3,978

    4,860

    4,397

    4,075

    3,602

    3,722

    4,137

    5b

    Tax voluntarily reported and paid ($m)

    43,570

    40,046

    41,315

    41,607

    37,943

    45,090

    51,275

    5c

    Theoretical liability ($m)

    47,548

    44,906

    45,712

    45,682

    41,545

    48,812

    55,412

    5d

    Net tax gap ($m)

    2,581

    2,867

    2,643

    2,252

    1,587

    1,699

    2,034

    5e

    Gross gap (%)

    8.4

    10.8

    9.6

    8.9

    8.7

    7.6

    7.5

    5f

    Net gap (%)

    5.4

    6.4

    5.8

    4.9

    3.8

    3.5

    3.7

    Note: Tax assured amounts are not used directly in the calculation, but feed into our calculations of unreported tax (Step 3) and non-detection (Step 4).

    Limitations

    Estimating the tax gap for large corporate groups is difficult and involves inherent uncertainty. Tax issues and tax laws are complex and contestable. Further, the estimates do not account for differences where there are alternative views on the appropriate interpretation of the tax law. In such circumstances, differences can exist between reasonably arguable positions presented by the ATO and by taxpayers.

    Non-detection is also challenging to estimate. We use tax assured data to improve estimates of non-detection where possible.

    The current methodology only provides an aggregated estimate of the large corporate groups tax gap. While this may allow generalised comparisons with other taxes, it does not measure relative risk between corporate groups or particular issues within this market.

    The gap estimate is a lagging measure, as compliance results take several years to flow through. This is due to the complexity of the tax issues in this population and the elapsed time associated with finalising our compliance activities.

    Find out about:

    Assumptions

    The assumptions used to construct this estimate are informed by actual data and expert opinion. The key assumptions are listed as follows.

    For those large corporate groups that we don't audit or review, we assume that: 

    • a certain degree of non-compliance with tax law occurs
    • the degree of non-compliance in these groups is less than those we do audit or review due to our risk detection approaches.

    For those large corporate groups that we do audit or review, we assume that: 

    • adjustments to their tax liabilities are representative of the value of non-compliance with tax law
    • we don't detect all instances of non-compliance
    • adjustments to their tax liabilities from completed audits and reviews are correct with the law at the time of estimation.

    For projected estimates, we assume that: 

    • past outcomes of audits, reviews, settlements and objections are reasonable representations of future outcomes
    • our tax assurance activities will continue to improve the accuracy our tax gap estimates as more tax assured data becomes available.

    Accounting for non-detection in the gap

    Not all errors are detected through audit and assurance activity. We account for this by applying a non-detection uplift to the unreported tax estimate.

    We apply different non-detection uplift rates depending on the level of assurance we have over the tax reported in each tax return. Where we have reviewed a tax return and have a high level of confidence in the amounts reported, we apply a lower non-detection uplift rate. We apply a higher non-detection uplift rate for tax returns that we have not reviewed.

    For the 2017–18 year, we estimated the impact of non-detected errors to be $1.1 billion.

    Updates and revisions to previous estimates

    Figure 2 displays the gross gap and net gap from our current model compared to our previous estimates and shows a downward trend.

    Figure 2: Current and previous large corporate groups income tax gap estimates, 2008–09 to 2017–18

    Figure 2: This graph is a representation of the previous and current net gap estimates provided with an overall downward trend. All data points are published in the data.gov file and outlined in Table 3.

    There have been no major changes made to the methodology since the last release of estimates in 2019. As such, the overall size and trend of the gap is similar to previous estimates. As standard practice, the estimates have been revised using updated data. The updated data provides additional information on the amount of tax assured, as well as the actual amount of amendments which reduces the reliance on earlier projections. The data used in Figure 2 is presented in Table 3 below.

    Table 3: Current and previous large corporate groups net tax gap estimates (percentage), 2008–09 to 2017–18

    Year published

    2008–09

    2009–10

    2010–11

    2011–12

    2012–13

    2013–14

    2014–15

    2015–16

    2016–17

    2017–18

    2017

    6.3%

    6.5%

    4.7%

    5.8%

    5.8%

    5.8%

    5.8%

    n/a

    n/a

    n/a

    2018

    n/a

    6.5%

    5.5%

    5.5%

    6.5%

    6.1%

    4.5%

    4.4%

    n/a

    n/a

    2019

    n/a

    n/a

    4.9%

    5.4%

    6.1%

    5.2%

    5.0%

    4.7%

    4.0%

    n/a

    2020

    n/a

    n/a

    n/a

    5.4%

    6.4%

    5.8%

    4.9%

    3.8%

    3.5%

    3.7%

    We will publish revisions to these results in future years as information becomes available. New information generally relates to later years and by including this we can reduce the uncertainty in the estimates and improve their reliability and credibility. Given the higher level of uncertainty with later year gap estimates, caution should be taken in extrapolating these results.

    Return to:

      Last modified: 19 Oct 2020QC 57641