We use a 4-step bottom-up random enquiry approach to estimate the individuals not in business tax gap. Random sampling methods are highly credible and best practice. They are commonly used by international jurisdictions to estimate tax gaps for this type of population.
Step 1: Estimate unreported amounts and extrapolate to population then apply estimate for people outside the system
In each year we draw on a bundled sample of up to 3 years from the random enquiry program. The bundled sample is split into 2 key groups of those who are:
- progressed to manual review
- verified.
We combine the incidence rates and averages from these 2 groups. We then extrapolate to the population of individuals not in business to estimate the unreported tax liability.
We also estimate the impact of non-registration or non-lodgment of people outside the system. This estimate draws on comparisons of Australian Bureau of Statistics Census of Population and Housing data to income tax return data to estimate the number of non-lodging individuals who are not in business. We then estimate a dollar impact, drawing on the random sample data to determine the final amount. We discuss this further in Limitations.
Step 2: Estimate for errors not detected
We apply an uplift to the unreported tax liability estimate to correct for errors not identified through the random enquiry program. Uplift factors are based on the midpoint of international ranges and account for non-detected amounts relating to:
- income misreporting
- deductions and other issues.
We also apply an uplift for non-detected amounts that relate to hidden wages, consistent with our wider program for wages. We discuss this further in Limitations.
Step 3: Estimate for non-pursuable debt
We add in the value of non-pursuable debt. This is debt that the Commissioner of Taxation has assessed as:
- not legally recoverable
- uneconomical to pursue
- unable to be pursued due to another Act.
Debt trends show that it takes upwards of 5 years for non-pursuable amounts to crystallise or be considered finalised in any one financial year. As a result, we add a provisional amount of non-pursuable debt to the actual amount recorded in the most recent 4 years, based on historical amounts.
Step 4: Consolidate the gap estimates
We calculate the gross gap by adding the unreported amounts from Steps 1 to 3. We calculate the net gap by subtracting the total amendment amount from the gross gap. Then we add the net gap to the tax paid to estimate the total theoretical liability. We derive both net and gross gap ratios by dividing the dollar amounts by the theoretical liability.
Summary of the estimation process
Table 5 provides a summary of each step of the estimation process and the results for each year.
Step |
Description |
2014–15 |
2015–16 |
2016–17 |
2017–18 |
2018–19 |
2019–20 |
|---|---|---|---|---|---|---|---|
1.1 |
Estimate unreported amounts and extrapolate to population (m) |
6,847 |
7,268 |
7,290 |
7,455 |
6,697 |
6,874 |
1.2 |
Apply estimate for people outside the system ($m) |
127 |
134 |
230 |
206 |
183 |
87 |
2.1 |
Apply estimate for non-detection (excluding hidden wages) ($m) |
179 |
220 |
362 |
531 |
525 |
414 |
2.2 |
Apply estimate for hidden wages ($m) |
1,666 |
1,740 |
1,803 |
1,954 |
2,004 |
2,095 |
3 |
Estimate for non-pursuable debt ($m) |
161 |
177 |
194 |
220 |
202 |
202 |
4.1 |
Estimate the gross gap (by adding together the results of Steps 1 to 3) ($m) |
8,890 |
9,539 |
9,880 |
10,366 |
9,611 |
9,673 |
4.2 |
Subtract compliance outcomes and voluntary disclosures ($m) |
809 |
751 |
895 |
711 |
575 |
643 |
4.3 |
Net gap ($m) |
8,171 |
8,788 |
8,984 |
9,655 |
9,036 |
9,030 |
4.4 |
Add tax paid ($m) |
119,184 |
126,095 |
130,426 |
140,849 |
144,360 |
151,498 |
4.5 |
Theoretical liability ($m) |
127,355 |
134,883 |
139,411 |
150,504 |
153,396 |
160,528 |
4.6 |
Gross gap (%) |
7.1 |
7.1 |
7.1 |
6.9 |
6.3 |
6.0 |
4.7 |
Net gap (%) |
6.4 |
6.5 |
6.4 |
6.4 |
5.9 |
5.6 |
Find out more about our overall methodology, data sources and analysis used for estimating our tax gap estimates.
Limitations
Limitations with estimating this tax gap include:
- The 2019-20 estimate uses 2 of the 3 finalised random enquiry program sample years. This will be updated in future estimates.
- The precision of the estimate is limited by the sample size. To maintain suitable confidence intervals over time, we use an ongoing bundled sample.
- To reduce compliance costs for the taxpayer, we applied materiality thresholds at the data-driven review stage. However, if a case develops into a manual review, all items in the tax return are investigated regardless of value.
- There is no independent data source that can provide a credible or reliable macroeconomic-based estimate, unlike for indirect taxes.
- The 2013–14 financial year sample does not account for the effect on offsets of adjustments to income and deductions. Where there are adjustments, the estimates reflect the change in tax on taxable income.
- A limitation of the random enquiry program and similar programs in other jurisdictions is uncertainty around the impact of the non-detection error. The enquiries we undertake do not discover the full extent of non-compliance.
Accounting for non-detection in the gap
Not all errors are detected through the random enquiry program. We account for these by applying a non-detection uplift to the unreported tax liability estimate.
The 3 sources of non-detection for this tax gap relate to:
- income misreporting
- deductions and other issues
- hidden wages.
The unreported tax liability is divided into the above elements, with an appropriate non-detection factor then applied to each portion.
Source of non-detection |
2014–15 |
2015–16 |
2016–17 |
2017–18 |
2018–19 |
2019–20 |
|---|---|---|---|---|---|---|
Income misreporting (excluding hidden wages) |
162 |
188 |
333 |
340 |
293 |
156 |
Deductions and other issues |
18 |
32 |
29 |
191 |
231 |
258 |
Hidden wages |
1,666 |
1,740 |
1,803 |
1,954 |
2,004 |
2,095 |
Total non-detection |
1,845 |
1,960 |
2,165 |
2,485 |
2,529 |
2,509 |
Accounting for the shadow economy
The shadow economy concerns economic activity not declared, which may be a result of attempts to avoid tax obligations. We account for the shadow economy in this tax gap by considering the impacts of:
- hidden wages
- people outside the system
- undisclosed business activity.
Element |
2014–15 |
2015–16 |
2016–17 |
2017–18 |
2018–19 |
2019–20 |
|---|---|---|---|---|---|---|
Hidden wages |
1,666 |
1,740 |
1,803 |
1,954 |
2,004 |
2,095 |
People outside the system |
127 |
134 |
230 |
206 |
183 |
87 |
Undisclosed business income |
0 |
0 |
0 |
0 |
0 |
0 |
Total shadow economy impact |
1,793 |
1,874 |
2,033 |
2,160 |
2,187 |
2,182 |
Confidence in the random sample findings
A confidence interval quantifies the precision of the estimate from a random sample relative to the true value from the population.
A 95% confidence level is considered industry best practice in terms of statistical analysis. It is the most used level by researchers, including His Majesty's Revenue and Customs in the UK in its equivalent tax gap program.
Using a 95% confidence level means we are 95% confident that the true value of the net gap for 2019–20 lies in the confidence interval 4.6% to 6.7%, or $7.2 billion to $10.8 billion. The upper and lower bounds of the 95% confidence intervals follow. The gap estimates we make public reflect the mid-point of the lower and upper bound estimates.
Figure 5: 95% confidence interval – upper and lower bound estimates – individuals not in business income tax gap, 2014–15 to 2019–20
The estimate from the random enquiry program is not the only component of this tax gap estimate. To establish the overall gap, we also draw on operational data for specific compliance risk areas. For example, we looked at the:
- failure by employers to withhold tax and report wage income of their employees
- non-lodgment of income tax returns
- non-payment of debts.
We combined the operational and random enquiry program findings to produce an overall gap estimate.
Updates and revisions to previous estimates
Each year we refresh our estimates in line with the annual report. Changes from previously published estimates occur for a variety of reasons, including:
- improvements in methodology
- revisions to data
- additional information becoming available.
This gap was first published in 2018 and has been revised 4 times annually. In 2019 we realigned the population and estimate to be consistent with the wider tax gap research program. All historically revised data points are marginally lower than the original estimates.
Figure 6 displays the net gap from our current model compared to all previously published estimates.
Figure 6: Current and previous individuals not in business income tax gap estimates, 2013–14 to 2019–20
The data used in Figure 6 is presented in Table 8 below.
Table 8: Summary of published net tax gap percentages for individuals not in business, 2013–14 to 2019–20
Gap release year |
2013–14 |
2014–15 |
2015–16 |
2016–17 |
2017–18 |
2018–19 |
2019–20 |
|---|---|---|---|---|---|---|---|
2021–22 |
n/a |
6.4% |
6.5% |
6.4% |
6.4% |
5.9% |
5.6% |
2020–21 |
5.4% |
6.1% |
6.2% |
6.1% |
5.9% |
5.6% |
n/a |
2019–20 |
5.6% |
6.1% |
6.2% |
6% |
5.6% |
n/a |
n/a |
2018–19 |
5.6% |
6.2% |
6.4% |
n/a |
n/a |
n/a |
n/a |
2017–18 |
5.8% |
6.4% |
n/a |
n/a |
n/a |
n/a |
n/a |