Methodology

6-step top-down method

We use a 6-step top-down approach to estimate the luxury car tax (LCT) gap. To derive the theoretical LCT payable in any year, our estimate draws on:

• motor vehicle registration data
• Vendor Field Analytical and Characterisation Technologies System (VFACTS)
• additional internal ATO data.

Due to the data quality issues in the unit record price information within the registration dataset and new registrations not adequately capturing the total volume of new cars sold which attract LCT, we have applied a clustering approach. Cars are first separated into groups, or 'clusters', based on similar attributes to produce price distributions of those cars within each cluster. We then derive the proportions of the price distributions above the LCT thresholds for all clusters and map them to the number of vehicle transactions from the VFACTS data that fall within those clusters. The prices and volumes are subsequently multiplied together and aggregated to produce an overall estimate of the theoretical tax liability. The more detailed steps are outlined below.

Step 1: Decode and standardise vehicle data

The Vehicle Identification Numbers (VINs) from registration data are decoded to obtain the correct vehicle information, such as:

• make and model configurations
• fuel consumption.

This ensures the naming conventions are consistent across vehicles and allows us to compare elements of the sales data. The formats and information reported in these data sets have different structures, which frequently require manual review to obtain the best match possible.

Step 2: Remove LCT-exempt vehicles and LCT from vehicle prices

We remove registration and transaction data associated with vehicle types not subject to LCT, such as:

• dealer registrations
• emergency and commercial vehicles
• registrations older than 2 years from the time of manufacture or importation.

We then remove the LCT components from the purchase prices to obtain the values of the vehicles (inclusive of GST).

Step 3: Develop vehicle clusters and price intervals

We determine vehicle clusters based on the manufacturer, number of cylinders and body type. Using this set of separating criteria should result in similarly valued cars, for deriving the price distributions of new cars by cluster based on the registration data. Our key assumption is that pricing is typically driven by vehicle performance and features.

Fuel-efficient and non-fuel-efficient cars have different thresholds above which LCT is payable. These can be different by year, so we separate them into clusters by year. This allows us to consistently determine the LCT payable for similar vehicle types.

For each cluster, we derive the representative value of vehicles exceeding the LCT thresholds. To address the issue of the representative value being skewed by high-value cars, the price observations of LCT-applicable cars above the LCT thresholds are split into 20 intervals for each cluster.

The representative value within each interval is constructed from the mid-point between the mean and the maximum of the value spread in each interval. Here we are assuming that the actual mean lies between the reported mean and the maximum of the reported values.

Step 4: Determine LCT payable for each interval

We estimate the LCT payable for each price interval within a cluster.

To obtain the values of vehicles that are subject to LCT for each interval within a cluster we:

1. Obtain the quantity sold in each cluster from VFACTS.
2. Multiply by the proportion of cars in the cluster that meet the relevant LCT threshold (giving the number of LCT-applicable cars in the cluster).
3. Divide by 20 (the number of intervals in the cluster) to give the number of LCT-applicable cars in each interval.
4. Determine the marginal value above the threshold by taking the difference between the representative value in Step 3 and the LCT threshold in each interval.
5. Multiply the number of LCT-applicable cars by the marginal value.
6. Remove the GST component by multiplying by 10/11.
7. Multiply by the LCT rate of 33% to obtain the corresponding LCT payable for all units sold in each price interval.

Step 5: Calculate total theoretical liability

The total theoretical liability is determined by aggregating the LCT payable for all price intervals, in all clusters.

Step 6: Calculate gross gap and net gap

The gross gap is the difference between the theoretical LCT liability and accrued LCT revenue excluding the compliance amounts.

The net gap is the residual gap amount after compliance amounts have been considered in the revenue base. We calculate the unreported amount by subtracting non-pursuable debt from the net gap amount.

Summary of the estimation process

Table 2 shows the:

• summary of each step of the estimation process
• results for each year.

Table 2: Summary of estimation process for the luxury car tax gap, 2016–17 to 2021–22

Step	Description	2016–17	2017–18	2018–19	2019–20	2020–21	2021-22
1-5	Theoretical tax liability ($m)	710	816	746	748	936	1,015
6.1	Less final tax reported ($m)	684	707	676	647	880	961
6.2	Equals final LCT liability not reported ($m)	26	109	70	101	56	54
6.3	Add non-pursuable debt ($m)	8.7	14.6	7.8	7.8	7.8	7.8
6.4	Equals net gap ($m)	35	124	78	109	64	61
6.5	Add compliance outcomes and taxpayer adjustments ($m)	8.2	21.0	12.4	6.5	7.3	13.2
6.6	Equals gross gap ($m)	43	145	90	115	71	75
6.7	Gross gap (%)	6.1%	17.7%	12.1%	15.4%	7.6%	7.4%
6.8	Net gap (%)	4.9%	15.1%	10.4%	14.6%	6.8%	6.1%

Find out more about our overall research methodology, data sources and analysis for creating our tax gap estimates.

Limitations

The following caveats and limitations apply when interpreting the LCT gap estimates:

• All vehicle data is mapped by a unique VIN for each vehicle. We match VINs to the information on the specifications of the vehicles based on the first 8 or 9 digits of the VINs rather than the entire 17 digits.
• Resource-intensive data manipulation is required to:
  • identify the LCT-applicable population by analysing over 1,000 models and makes of cars to determine an estimated purchase price (or range) for each new or imported vehicle
  • determine fuel-efficient LCT vehicles by combining the volume of sales data from VFACTS and registration data
  • map line-by-line registration data to the semi-aggregated VFACTS data — due to inconsistencies in the data formats and information reported, this requires extensive manual reviews to find the best match available.
• Due to some data quality issues, some vehicles may be incorrectly categorised as non-fuel-efficient (or fuel-efficient) or misclassified to a cluster.
• Overall, the estimates can be sensitive to the clustering method applied. There is an element of judgment when grouping the cars based on their likeness.
• At this stage we are uncertain on the shadow economy impacts. More work needs to be done to isolate these amounts.

Updates 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.

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

Figure 2: Comparison of previously published estimates – LCT gap

This data is presented in Table 3 below.

Table 3: Current and previous luxury car tax net gap estimates (percentage), 2009–10 to 2021–22

Program year	2009–10	2010–11	2011–12	2012–13	2013–14	2014–15	2015–16	2016–17	2017–18	2018–19	2019–20	2020–21	2021–22
2024	n/a	n/a	n/a	n/a	n/a	n/a	n/a	4.9	15.1	10.4	14.6	6.8	6.1
2023	n/a	n/a	n/a	n/a	n/a	n/a	12.2	5.1	15.8	10.6	14.9	7.7	n/a
2022	n/a	n/a	n/a	n/a	n/a	6.9	12.2	7.0	8.6	7.9	3.3	n/a	n/a
2021	n/a	n/a	n/a	n/a	8.1	3.4	10.1	5.8	7.8	9.0	n/a	n/a	n/a
2020	n/a	n/a	n/a	n/a	8.1	3.4	10.1	5.8	7.8	n/a	n/a	n/a	n/a
2016	3.9	5.8	4.6	5.1	4.7	5.2	n/a	n/a	n/a	n/a	n/a	n/a	n/a
2015	4.1	4.3	4.1	4.3	3.3	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a
2011	4.9	5.2	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a