• ## Methodology

The PAYG withholding gap estimate is derived through applying a top-down method. We use five-steps in applying this method to estimate the gap. These steps are expanded on below followed by a summary of the overall estimate:

### Step 1: Estimate salary and wages subject to withholding

To calculate the PAYG withholding gap, we start with external data on Compensation of Employees – salary and wages published by the Australian Bureau of Statistics (ABS) in its quarterly and annual Australian National Accounts. We adjust the data in the following way:

• We apply an uplift factor of 1.2% to account for the shadow economy. In arriving at this uplift factor, we considered comparable international uplifts and conducted an analysis of the large amount of data from our interactions with employers and our compliance activities that focus on the shadow economy.
• We subtract benefits and payments not subject to withholding. This includes payments in kind (fringe benefits tax), salary sacrifice into super, employee share schemes, and defence exempt income.
• The resulting amount is assumed to represent salary and wages subject to withholding.

### Step 2: Calculate the theoretical PAYG withholding amount

In this step, we multiply the amount calculated in Step 1 by the average rate of withholding to calculate the withholding amount that should be reported and remitted to us.

The average rate of tax withheld is calculated from tax return data for each year. This is the amount of tax withheld, from salary and wage income and other items such as lump sum payments made by employers to their employees and allowances over the total salary and wages type income.

This provides us with the theoretical PAYG withholding amount that should be reported and remitted. For 2017–18, we had to make a provision to the theoretical liability amount to be consistent with a gross gap of 3.3%, the same as that of 2016–17.

### Step 3: Calculate PAYG withholding received by the ATO

In Step 3, total PAYG withholding received is calculated from our systems data. It is calculated on an accrual basis. Non-resident dividends, interest and royalty withholding tax amounts are excluded.

We subtract amounts not subject to withholding. These include personal services income for sole traders, tax withheld where an Australian business number (ABN) is not quoted, voluntary agreements, benefits and pensions. We are then left with PAYG withholding received minus excluded amounts.

### Step 4: Calculate the PAYG withholding net gap and gross gap

In Step 4, we subtract PAYG withholding received (calculated in Step 3) from the theoretical PAYG withholding amount (calculated in Step 2). We then add non-pursuable debt amounts (PAYG withholding liabilities which have been written-off after being deemed as irrecoverable at law or not economical to pursue) to determine the net gap.

Finally, we add compliance outcomes and taxpayer self-adjustments to the PAYG withholding net gap to determine the PAYG withholding gross gap.

For 2017–18 we continue to see the divergence between our data and ABS data that results in a negative gap. For the purpose of this year's estimate we apply a projection pending a revision by ABS to their data.

### Summary of the estimation process

Table 2 provides a summary of each step of the estimation process and the results for each year.

Table 2: Summary of estimation process (value) for PAYG withholding gap

Step

Description

2012–13

2013–14

2014–15

2015–16

2016–17

2017–18

1.1

Salary and wages (\$m)

664,879

686,956

706,091

727,219

745,665

780,121

1.2

7,979

8,243

8,473

8,727

8,948

9,361

1.3

Earnings not subject to withholding (\$m)

26,809

26,965

28,993

29,740

29,979

30,010

1.4

Salary and wages subject to withholding (A) (\$m)

646,049

668,235

685,571

706,205

724,634

759,472

2.1

Average rate of withholding (B)

0.237

0.239

0.247

0.249

0.250

0.252

2.2

Multiply (A) and (B) to calculate theoretical liability (C) (\$m)

152,886

159,971

169,285

175,912

181,161

191,759

2.3

Make a projection for 2017-18 theoretical liability

152,886

159,971

169,285

175,912

181,161

194,057

3.1

150,196

157,399

166,829

174,196

180,397

193,020

3.2

Subtract excluded amounts (\$m)

1,355

1,402

1,622

1,716

1,795

1,878

3.3

PAYG withholding received minus excluded amounts (D) (\$m)

148,841

155,997

165,207

172,480

178,602

191,142

4.1

Non-pursuable debt (E) (\$m)

563

591

570

570

570

570

4.2

(C) minus (D) plus (E) to equal net gap (F) (\$m)

4,607

4,565

4,648

4,002

3,128

3,484

4.3

Add amendments to net gap (F) (\$m)

1,334

1,431

2,162

2,367

2,908

2,981

4.4

Equals gross gap (\$m)

5,941

5,996

6,810

6,369

6,036

6,466

4.5

Gross gap (%)

3.9

3.7

4.0

3.6

3.3

3.3

4.6

Net gap (%)

3.0

2.9

2.7

2.3

1.7

1.8

### Limitations

Our data for salary and wages from tax returns does not have the same definition as the salary and wages series produced by the ABS. While we make adjustments to improve consistency, there are some items in the ABS data series we cannot quantify. The main variances are:

• Individuals earning under the tax-free threshold are not required to lodge a tax return. While a significant number of these people do lodge tax returns, some wage and salary data will be missing. This will affect the average tax rate we apply.
• We do not have data on changes in provision for future entitlements, such as long service leave.
• While we have tax expenditure statement estimates for tax-exempt Australian Defence Force payments and allowances, this does not include the value of food, clothing and travel.
• We are uncertain about the number of individuals under voluntary withholding agreements or labour hire arrangements where they are classified as employees.

We use ABS National Accounts data to estimate theoretical salary and wages amounts. There are limitations with this approach, which means there will be a margin of error with the data. The limitations include:

• sampling and non-sampling errors
• National Accounts data is often subject to revision, which can vary historical trend results. Material revisions to this data will result in changes to the estimated PAYG withholding gap
• non-measurement of the underground economy in relation to Compensation of Employees due to lack of sufficient data. We have added an uplift factor of 1.2% to account for this and ensure the gap is not underestimated.
• As highlighted in the Black Economy Taskforce final report it is likely the uplift for hidden wages understates the true nature of the economic impact on undisclosed wages. We will be reviewing this estimate in the future with an expectation of an upwards revision.

Some of the key assumptions of the methodology are:

• When we remove identified income not subject to withholding from National Accounts data, the resulting amount represents salary and wages subject to withholding.
• The average rate of withholding we apply to this data is appropriate.

#### Updates and revisions to previous estimates

Each year we refresh our estimates in line with our annual report. Changes from previously published estimates occur for a variety of reasons, including:

• improvements in methodology
• revisions to data

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

Figure 2: Current and previous PAYG withholding gap estimates, 2008–09 to 2017–18

Figure 2: This graph provides a visual representation of the published gaps.

The data is presented in Table 3 below.

Table 3: Current and previous PAYG withholding 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

2015

2.0%

2.1%

2.2%

2.1%

n/a

n/a

n/a

n/a

n/a

n/a

2016

2.5%

2.5%

2.9%

3.0%

2.1%

1.4%

n/a

n/a

n/a

n/a

2017

n/a

3.7%

3.9%

4.1%

2.9%

2.7%

1.8%

n/a

n/a

n/a

2018

n/a

n/a

4.3%

4.1%

2.9%

2.7%

2.6%

1.9%

n/a

n/a

2019

n/a

n/a

n/a

4.2%

3.0%

2.8%

2.7%

2.2%

1.7%

n/a

2020

n/a

n/a

n/a

n/a

3.0%

2.9%

2.7%

2.3%

1.7%

1.8%