Official Statistics

Background documentation

Published 30 September 2020

1. Income Tax and NIC liability for company cars and car fuel benefit

Directors of companies and employees are liable to pay income tax on the value of their company cars and car fuel benefits. Their employers (or in certain cases other third parties who provide benefits in kind) are liable to pay Class 1A NICs on these benefits. Other benefits may be subject to Class 1 NICs. The taxable value of a benefit depends on its type.

Company cars

Since April 2002, the taxable value of a company car has been based on the car’s list price and appropriate percentage (linked to the car’s CO2 emissions).

Car fuel benefit

Since April 2003, the taxable value of employer-provided fuel has been calculated using the fuel benefit multiplier and the appropriate percentage used to calculate the car benefit (linked to the car’s CO2 emissions).

Further details

Further details on the calculation of taxable values for company cars, car fuel benefit, and other benefits in kind can be found in HMRC Booklet 480 (Expenses and Benefits).

In general, the taxable value of benefits in kind may be reduced by contributions paid by the employee for the benefit in kind. The final value of the benefit for tax purposes will be used to assess the Class 1A NICs that the employer (or third party) is liable to pay. Taxable benefits not already subject to Class 1 NICs are liable for Class 1A NICs unless they are exempt.

Further details about the taxation and NICs treatment of benefits in kind and expense payments can be found on the HMRC and gov.uk websites using the following links:

• Employment Income Manual
• Expenses and Benefits: A tax guide (Booklet 480)
• Class 1A National Insurance contributions on benefits in kind: A guide for employers (Booklet CWG5)
• Expenses and benefits for employers

2. Data sources and methodology

This section describes the data sources and methods used to compile statistics for the taxable value and number of recipients of company cars and car fuel benefit, as well as the tax and NIC liability, split by different dimensions shown in Tables 4.1, 4.3, 4.4, and 4.5 of this release.

2.1 Data sources

The published estimates are based on the data collected from P11D forms, RTI submissions, and the Survey of Personal Incomes (SPI).

P11D forms

At the end of each tax year, employers must provide HMRC with details of taxable expenses and employment benefits provided to directors and certain employees. Employers are required to complete a separate P11D form for each employee whose expenses and benefits they report. P11D data are captured on the HMRC Employer Compliance System (ECS).

RTI submissions

From April 2016 the requirement to report certain benefits in kind on the P11D form has been removed for those employers who register to deduct the tax due on those benefits in kind directly from payroll. From April 2017, employers who payroll company cars or car fuel benefits could voluntarily report details as part of their regular PAYE returns on RTI (Real Time Information). From April 2018, reporting these details on RTI is a requirement for employers who payroll company cars or car fuel benefits.

Survey of Personal Incomes (SPI)

The SPI is a sample survey derived from data held by HMRC on persons who have been in contact with HMRC over the tax year through the PAYE, SA or repayment claims processes. The survey consists of a systematic random sample of individuals each tax year. For each individual in the sample, SPI includes information on incomes assessable to income tax for the tax year, together with some basic information on individual characteristics, for example age and gender.

The SPI sample size has increased over time to around 700,000. It is made up of three separate samples drawn from three different HMRC administrative systems. SPI datasets are available for public use via the UK Data Archive at Essex University (registration required).

2.2 Methodology

Following the end of each tax year an individual level dataset, containing all benefits in kind and taxable expenses that have been reported on a P11D return, is compiled.

For the years 2017 to 2018 and 2018 to 2019 within this publication, the P11D data is supplemented with RTI data to account for car and car fuel benefits that were not reported on the P11D form due to payrolling.

The number of recipients and the amount of each car and car fuel benefit can be obtained from the combined P11D and RTI data, and are reported in Tables 4.1, 4.3, 4.4, and 4.5.

A data cleaning exercise is performed to make sure the source data are fit for purpose. Some company cars were reported on both P11Ds and in RTI data. For these cases, an exercise is carried out to identify the duplicated company cars and remove one version.

The Survey of Personal Incomes (SPI) is then merged onto the combined P11D and RTI data to obtain information relating to total income and tax liability as this is not currently held on the combined data. In 2017 to 2018, around 25,000 individuals in the selected sample received company cars or car fuel benefits reportable in these tables.

Calculating Income Tax and National Insurance contributions

An average tax rate per pound of benefit is calculated from the combined P11D and RTI data merged with the SPI sample. The tax and NIC liabilities in respect of benefits in kind for an individual are calculated as the difference between liabilities arising with and without the value of benefits included.

The relevant tax rates applicable for each tax year on earnings, savings and dividends have been used to calculate the tax liabilities, taking into account various deductions which reduce an individual’s tax liability such as the personal allowance and reliefs.

The difference between the tax liabilities including and excluding benefits is summed over all SPI sample cases (with grossing factors) and then divided by the summed total benefit amount to give an average tax rate per pound of benefit. This average tax rate (for each benefit) is then applied to the total taxable value derived from the combined P11D and RTI data to obtain an overall tax liability estimate.

This approach effectively considers the benefit in kind as the top element of income, and consequently the tax liability reflects the taxpayer’s highest rate of tax. An alternative approach would have been to attribute the amount of tax to the benefits in kind based on the proportion of the benefit in kind value over total income. This would have resulted in lower average tax rates compared to the approach adopted with a considerable impact on tax liabilities calculated.

The methodology adopted should be better suited to most users’ needs. The tax liability shown represents the tax figure which would be affected by any changes to rules for determining taxable value of benefits, or by any change to take-up of benefits in kind.

A corresponding methodology was adopted for the NIC liability. For Class 1A NICs paid on company cars and car fuel benefits, the rates are 12.8% of the value of the benefit (employer only) for years up to 2010 to 2011 and 13.8% for years from 2011 to 2012. This is then multiplied by the total value of the benefits from the combined P11D and RTI data to produce a value for National Insurance contribution liability.

Table 4.1 reports the number of recipients and the amount of car and car fuel benefits along with their tax and National Insurance contributions.

Definition of total income

The total amount of car and car fuel benefits split by total income, inclusive of the benefit amount, are reported in Table 4.3. Total income includes total earned income as well as total investment income; not just earnings from employment. Total earned income includes state benefits as well as earnings from employment such as taxable incapacity benefit, other taxable social security benefits and jobseeker’s allowance payments. Gross profits assessable for all sources of self-employment income, taxable pay on termination of employment, and pensions are also included. Total investment income would include net income from UK and overseas property, net interest from UK banks, building societies and other deposit takers, dividends from shares in UK companies and unit trusts, gains on life policies with tax treated as paid, gross interest paid, and any other interest & dividends from UK and overseas savings & investments charged at 20%.

Estimating breakdown of totals according to total income

Table 4.3 provides a breakdown of total values according to range of total income.

Estimates for each range are found from the SPI sample, as the combined P11D and RTI data do not contain all details of income. The estimates are grossed up using standard SPI grossing factors, but a further adjustment factor is then applied in order that the total taxable benefit value for all cases is equal to the total figure known from the full data.

The same adjustment factor is used for both benefit value and number of recipients, with the result that the number of recipients reported by income band will not sum exactly to the total number of recipients from the full data.

It is not possible to adjust the number of recipients to the value in the full data by a different factor than that used for the amount of benefit without distorting estimates of average benefit value.

Calculation of number of car recipients split by CO2 emission

Table 4.4 reports all car benefits split by the CO2 emission of each benefit. However, since a recipient can have more than one car benefit in the tax year, it was necessary to apportion that recipient between their multiple car benefits, so that they would not be counted more than once. A recipient was therefore apportioned between their car benefits according to the length of time they had each benefit.

Where a recipient only had a car benefit for part of the year, the car benefit, tax and NIC values in the table reflect the part year. However the individual still counts as one person for the purposes of the recipients column. Where a recipient had multiple car benefits at the same time so that they had the equivalent of one benefit for greater than a year, this was scaled down to a year, and the recipient apportioned between them.

Imputation of CO2 values and fuel type

CO2 values are reported on the P11D and RTI data where an employer reports a car benefit. Sometimes these are missing or are implausible values. Where other supporting data can help impute the missing or implausible value, this has been done prior to producing the figures in Table 4.4. Less than 2% of car benefit values were imputed for 2017 to 2018.

After a cleaning exercise, the average tax rate per pound of benefit was calculated for each appropriate percentage (i.e. different bands of CO2 emissions). This is because there is evidence that higher rate taxpayers tend to have higher emitting cars. Where the appropriate percentage could not be identified, an average tax rate per pound of benefit was used from all car benefits. This was multiplied by the amount of each car benefit to provide a tax liability. The National Insurance Contribution was 13.8% of the value of the benefit.

Due to sampling error, the sum of tax liabilities across CO2 bands would not equal the total sum calculated for Table 4.1, therefore an adjustment factor is used to make the total tax liability match that calculated for all car benefits. The same average tax rates used across CO2 emissions were applied when grouping by the type of fuel in Table 4.4. This was adjusted to the total sum calculated for Table 4.1 in the same way.

Table 4.4 also includes some information on average CO2 emissions (including a breakdown by fuel type).

Projections to 2018 to 2019

The SPI data have been projected forward to 2018 to 2019 based on the individuals selected for the 2017 to 2018 sample.

The number of recipients and taxable value of the benefit for 2018 to 2019 are obtained from P11D and RTI data for 2018 to 2019 and are not based on projections. The projected SPI for 2018 to 2019 is then merged onto the P11D and RTI data for 2018 to 2019 to obtain information relating to total income and tax liability. The average tax rate per pound is then calculated in the same way as previously described. This average tax rate and the Class 1A NIC rate are then applied to the combined P11D and RTI data for 2018 to 2019.

3. Quality

Users should be aware of various issues affecting the completeness and accuracy of the tables. A Quality Report covering these statistics is available alongside these statistics.

This report assesses the statistics against different attributes of quality such as relevance, accuracy and reliability, timeliness and punctuality, accessibility and clarity, and coherence and comparability.

The following section contains more detailed information about two particular aspects of quality: completeness and accuracy.

3.1 Completeness

Some company cars are not included in the source data used to produce the statistics and are therefore excluded from the tables. They are:

Unreported or uncaptured payrolled company cars.

These are company cars where the employer has deducted tax through their payroll system.

Before April 2016 a small number of cars were subject to ‘informal payrolling’. In such cases P11Ds were submitted to HMRC but not captured on ECS.

From April 2016, employers have been able to register for ‘voluntary payrolling’, removing any requirement to submit a P11D return when deducting tax through the payroll system. The number of schemes payrolling company cars has increased year on year since 2016 to 2017.

This has created incompleteness in HMRC’s data in the following ways:

• in the first year of voluntary payrolling 2016 to 2017, there was no system for reporting payrolled company cars

• in 2017 to 2018, payrolled company cars could be reported in PAYE submissions via Real-Time Information (RTI) but this was on a voluntary basis

• in 2018 to 2019 reporting of payrolled company cars became mandatory but analysis suggests there is significant under-reporting

It is difficult to produce accurate estimates of the number of unreported company cars. However initial analysis suggests that they account for a large part of the reduction in reported cars observed since 2015 to 2016.

Company cars on late arriving P11D forms

P11D forms are due to be filed in the July following the end of the tax year. The data used in the published tables come from extracts from live data and any P11D forms processed after the extract dates will not be reported in the tables. These P11Ds will either be new cases or changes to existing cases.

Estimated “late-arriving factors” for the data used in these statistical tables, representing the additional proportion of benefit value likely to arrive after the extract date, are:

• 0.2% for tax years up to 2017 to 2018

• 0.8% for tax year 2018 to 2019

Zero benefit values

In a relatively small number of cases where the reported taxable benefit value of the company car was zero (for example where an employee contribution matched or exceeded the taxable benefit value) the cases were discarded.

3.2 Accuracy

The following is a list of the key issues with a bearing on the accuracy of the figures provided.

• there will be sampling error for those figures which depend on the Survey of Personal Incomes (tax and NIC liability estimates, and breakdowns by total income)
• errors in the reported data may result in a small amount of inaccuracy. We undertake data cleaning, including the imputation exercise referred to in Annex B, as a way of reducing inaccuracy