Official Statistics

Quality report: Scottish VAT Assignment

Updated 28 September 2023

The latest release was published 09:30 28 September 2023. The next release will be published at a date to be confirmed.

1. Contact

• organisation unit: Customer Strategy and Tax Design - Knowledge, Analysis and Intelligence (KAI)

• name: R Nandra

• function: Statistician - Indirect Taxes, Customs & Coordination

• email: modelscottishvatassignment@hmrc.gov.uk

2. Statistical presentation

2.1 Data description

This Experimental Statistics publication on Scottish Value Added Tax (VAT) Assignment (VA) presents the share of the first 10 percentage points of standard rated VAT and 2.5 percentage points of reduced rated VAT raised in Scotland as an annual proportion of UK VAT receipts.

2.2 Sector coverage

The share has been estimated using the Scottish VA model, jointly developed by officials from UK and Scottish Governments. The VA share calculation is based on a detailed model which considers how much VAT Scotland would raise if it were a separate jurisdiction for VAT purposes considering expenditure in Scotland and the rest of the UK on different types of goods and services.

There are various data sources providing independent expenditure data which are used in the model to estimate consumption in Scotland and other UK regions.

There are several areas of expenditure which contribute to UK VAT, with the largest contribution coming from household spending. The VA model is made up of 5 main spending components and several adjustment components. Each spending component represents an area of expenditure where the final VAT liability is generated.

The 5 expenditure components that make up the VA model are as follows:

• household
• charities
• central government
• exempt
• housing (expenditure on construction and repairs that are not included in household)

There are also adjustments to ensure VAT rules are applied correctly.

2.3 Statistical concepts and other definitions

UK VAT receipts

The amount of VAT collected by HMRC.

VAT liabilities

The amount of VAT that taxpayers must pay to HMRC. VAT liabilities are accrued in the financial year where the end date of the company’s accounting period lands.

Calendar year

The statistics are aggregated into calendar years. A calendar year stretches from 1st January until 31st December.

Financial year

The calendar year statistics are applied to VAT receipts for their corresponding financial year. A financial year stretches from 01 April until 31 March the following calendar year.

VAT rates

There are 3 different VAT rates in the UK, with different goods and services having different rates.

Most goods and services are taxed at the standard rate, which is currently 20%. There is also a reduced rate (currently 5%) and a zero rate (0%) for goods which are eligible for VAT but not currently charged. The VA model accounts for and is unaffected by any changes in VAT rates.

VAT exempt

Some goods are exempt from VAT entirely and therefore not charged. For these goods, input VAT cannot be reclaimed on any inputs for their production, whereas it can for zero rated goods. Income from these goods also does not contribute towards reaching the turnover threshold for mandatory VAT registration.

VAT registration

Businesses with a turnover above a certain threshold, not including turnover from VAT exempt goods, are required to register for VAT and submit regular VAT returns and payments. Businesses with a turnover below this threshold are not required to register and therefore do not need to pay VAT on goods sold.

Some businesses below this threshold do decide to register for VAT, as it allows them to reclaim VAT paid on goods purchased for the running of the business.

Standard rate equivalent (SRE) expenditure

SRE expenditure is where expenditure liable to VAT is converted to a standard rate equivalent. This means the expenditure of an item is calculated as if it were standard rated, that is, as if VAT were 20%.

For example, if net expenditure (excluding VAT) of £10.00 is liable to a reduced rate of VAT (5%), gross expenditure would be £10.50 (net expenditure + VAT of £0.50). The standard rate equivalent would be a gross amount that includes the same level VAT, in this case £0.50 at the standard rate of 20%.

The standard rate equivalent would be net expenditure of £2.50, because with a standard rate applied (20%), the VAT would be £0.50, and gross expenditure would be £3.00.

SRE expenditure presented in this publication is all on a gross expenditure basis (meaning it is inclusive of VAT).

2.4 Statistical unit

The unit in the statistics is the estimated Scottish VA share. There are also statistics (in £ million and percentages) including UK total VAT receipts, the illustrative VA share, revisions to the back series, statistical uncertainty estimates, and breakdowns by Scottish VA components.

2.5 Reference area

The model calculates the VAT assigned to the area of Scotland, but uses data from the rest of the United Kingdom (England, Wales, and Northern Ireland) to establish the proportion of UK VAT receipts assigned to Scotland.

For the purpose of the VA model, Scotland is defined in line with the UK’s International Territorial Level 1. This defines Scotland as its 32 land-based Local Authority areas and does not include territorial waters or the adjacent waters which form the Scottish zone of the UK exclusive economy zone.

2.6 Time coverage

The statistics include historical estimates from 2011 until the latest year for which data is available to produce an estimate.

3. Statistical processing

3.1 Source data

Household consumption

The main data source for regional proportions of household consumption is the Office for National Statistics (ONS) Living Costs and Food (LCF) Survey. This makes up approximately 70% of the household sector and is supplemented by many other minor data sources, including:

• ONS Opinions and Lifestyles Survey

• household energy costs published by the Department for Energy Security and Net Zero

• road fuel published by the Department for Energy Security and Net Zero

These regional proportions are applied to ONS Consumer Trends data, which provide overall UK expenditure in the full range of household expenditure categories.

Government

The main data sources for the Government and Exempt sectors are the ONS Annual Survey of Hours and Earnings (ASHE) and Business Register and Employment Survey (BRES), which provide average earnings and number of employees respectively, segmented across region and occupation.

Tourism

Tourism is split into 2 parts; Foreign tourism (data sourced from the International Passenger Survey)and domestic tourism (data sourced from Visit Britain, Visit Scotland and Welsh Government).

VAT Returns

VAT business and returns data is sourced from HM Revenue & Custom’s (HMRC’s) VAT system. This data is gathered through business registration for VAT and subsequent returns submitted to HMRC.

3.2 Frequency of data collection

Most of the data sources used are compiled on an annual basis, though some are published quarterly or monthly, but used as an annual aggregation.

VAT returns are collected continuously, and the data source refreshed daily.

3.3 Data validation

The majority of the survey data which is used for the VAT assignment model is published and has therefore already been checked and validated by statistics producers. The largest component of the model – the household section – uses LCF Survey data which undergoes validation by the ONS. More information on LCF Survey data validation can be found in the LCF technical report and LCF Quality and Methodology Information (QMI).

Each new estimate is compared with the previous annual estimate alongside changes to the back series. Any changes are investigated by looking at individual model components and SRE expenditure.

When inputting data into the model, quality assurance checks are performed to ensure the accuracy of the data inputs. The model itself is also quality assured, and has built-in quality assurance checks to flag discrepancies.

TheVA estimate results are compared to population proportions and GDP which are partly comparable independent estimates, so that any differences are understood.

The VA model has several assumptions where data may be missing or data needs to be smoothed, such as when there are small sample sizes.

3.4 Data compilation

Aggregating data

Data are aggregated to the regional level so that the Scottish share of VAT can be calculated, and data sources which produce monthly or quarterly outturns are aggregated to annual outturns.

3.5 Adjustment

VAT deferral schemes

In 2020 during the COVID-19 pandemic, a VAT deferral scheme was introduced which allowed VAT-registered traders to defer VAT payments from 20 March 2020 to 30 June 2020 until 31 March 2021. A further scheme was introduced which allowed customers to repay VAT due from 20 March 2020 to 30 June 2020 in installments up until March 2022.

A manual adjustment has been made to account for the lag in VAT payments arising from the deferral scheme for the years 2020 and 2021. The adjustment reassigns VAT receipts that would have arisen in 2020 but had their payment deferred until 2021, to the year 2020.

4. Quality management

4.1 Quality assurance

All official statistics produced by KAI must meet the standards in the Code of Practice for Statistics produced by the UK Statistics Authority and all analysts adhere to best practice as set out in the ‘Quality’ pillar.

Analytical Quality Assurance describes the arrangements and procedures put in place to ensure analytical outputs are error free and fit-for-purpose.

At the start of a project, during the planning stage, analysts and managers make a risk-based decision on the level of QA required.

Analysts and managers construct a plan for all the QA tasks that need to be completed, along with documentation on how each of the tasks are to be carried out and turn this list into a QA checklist specific to the project.

Analysts carry out the QA tasks, update the checklist, and pass on to the Senior Responsible Officer for review and eventual sign off.

4.2 Quality assessment

The QA for this project adheres to the framework described in ‘4.1 Quality assurance’ and the specific procedures undertaken are as follows:

Stage 1 – building and populating a model

Each individual piece of data is extracted and formatted using the most appropriate software and in line with good practice guidance. Data inputs are checked to ensure they are fit-for-purpose by reviewing available documentation and, where possible, through direct contact with data suppliers. The input data and outputs for the model are audited by someone other than the lead analyst – checking code and methodology.

Stage 2 – running and testing the model/code

The outputs in the summary model are compared with those produced in previous years and differences understood and determined to be genuine. Differences have been investigated by looking at model components and SRE expenditure. Results are compared with limited comparable independent estimates, and differences understood. For example, VAT assignment outputs are compared to population proportions and GDP. Results are determined to be explainable and in line with expectations.

Stage 3 – drafting the final output

Checks are completed to ensure internal consistency (for example, totals equal the sum of the components). The final outputs are independently proofread and checked.

5. Relevance

5.1 User needs

This analysis is likely to be of interest to the following users:

• UK Government and Scottish Government (inlcuding MPs and MSPs) for policy making and policy decision

• Office for Budget Responsibility and the Scottish Fiscal Commission for use in their independent economic forecasting

5.2 User satisfaction

KAI acts on user feedback, wherever possible and we are always open to ideas for new analysis to meet changing user requirements.

6. Accuracy and reliability

6.1 Overall accuracy

This analysis is based on both administrative and survey data. For administrative data, accuracy is addressed by eliminating non-sampling errors as much as possible through adherence to the quality assurance framework. For survey data, each source has been identified as the best data for its purpose.

The VA model uses well established surveys where producers perform their own quality assurance to assure accuracy as much as possible. The LCF Survey is the primary data source used to calculate the Scottish share of UK household expenditure in the VA model and accounts for around 85% of the regionalisation of expenditure for the entire VA model outturn.

Due to its significance, HMRC and Scottish Government have jointly sponsored a boost to the Scottish sample size of the LCF Survey from the 2017 to 2018 financial year onwards to increase the accuracy of the regional splits for household expenditure.

The potential sources of error include:

• sampling error from the use of survey data

• human or software error when inputting new data or running calculations

• systematic errors where the modelling assumptions, errors in the data or missing data lead to estimates that are consistently either too low or too high

Due to COVID-19 affecting several data sources (in particular surveys) in 2020 and 2021, there is likely to be more error introduced into the model in these years due to an increase in methodology adaptations.

6.2 Sampling error

Many data sources are used in the VA model to calculate regional SRE expenditure. The most significant data sources that have large impacts on the model have been evalated. The results formed part of the 2019 VAT assignment publication.

Statistical uncertainty in the VAT assignment model

Preliminary analysis of the most significant data sources for allocating expenditure by region has been undertaken for the LCF Survey and the combined inputs of the Annual Survey of Hours and Earnings (ASHE) and the Business Register and Employment Survey (BRES).

Together these elements drive 80% of the regionalisation of total VAT liabilities. This analysis aims to calculate 95% confidence intervals for the VAT assigned to Scotland and initial results suggest these could be in the region of ±£135 million (±4.0%) of the assigned VAT driven by LCF Survey data, and ±£80 million (±7.0%) of the assigned VAT driven by ASHE and BRES data together.

The complexity of the VA model means there are limitations to the methods used and, as such, it is possible these initial findings have been underestimated.

Estimates are based on the 2017 VA share of £5,430 million and its associated inputs, as this was the latest data available at the time of the analysis. The LCF Survey drove £3,340 million of this share and ASHE and BRES inputs drove £1,110 million.

Table 1: Summary of initial statistical uncertainty estimates for the LCF Survey, ASHE and BRES data inputs

VA Model data source	Share of assigned VAT	VAT assigned using this data (£ million)	Estimated 95% confidence interval (£ million)	Confidence interval as share of assigned VAT	Confidence interval as a share of the final Scottish VA share
LCF Survey	61%	3,340	±135	±4.0	±2.5%
ASHE and BRES	21%	1,110	±80	±7.0	± 1.5%

6.3 Non-sampling error

Coverage error

The main sources of coverage error would be from survey data. HMRC and Scottish Government jointly sponsored a boost to the Scottish sample size for the LCF Survey as the primary source for regionalising UK expenditure. This included an increase of coverage for Scotland geographical areas, in order to minimise this type of error.

Measurement error

The main sources of measurement error could be categorised as respondent errors and include the following:

• for administrative data, companies may make errors entering their information onto tax return forms, whether this is done on paper or electronically

• for survey data, survey respondents may make errors providing their details and expenditure information

There is a risk that errors involving very large expenditure amounts may distort the overall statistics. To mitigate against this, checks are conducted by comparing results to the historic series, previous versions and alternative measures.

Missing data sources has been a particular issue in the VA model for 2020 and 2021 due to COVID-19 impacting on survey data. For missing data sources, an approach has been developed by HMRC and agreed with the Scottish Government to estimate expenditure values. Some of these approaches act as an interim until data is available, while others are permanent as the data source for those years will not be available in the future.

Forecasting for missing data sources is based on historical data and trends. Where necessary, forecasts have been adapted to follow expected trends using available data as a guide. For example, forecasts produced for 2020 and 2021 have had COVID-19 implications considered when developing a methodology.

Nonresponse error

For administrative data, nonresponse error should be mitigated due to the timing of the VA publication. The estimate is released around 21 months following the end of a calendar year. For example, the 2019 VA estimate was published in September 2021. As such, almost all companies will have completed returns and assessments for the relevant period.

For survey data, there is a risk of nonresponse error. Nonresponse for surveys used within the VA model could result in regional expenditure being under or overrepresented and skew the final outturn. The ONS acknowledges that falling response rates over recent years is a problem and have instigated various initiatives to address this. For more information on this, see the LCF Technical report.

The LCF Survey response rate for the 2021 to 2022 financial year was 27%, rising from 24% in 2020 to 2021.

Processing error

It is possible that errors exist in the programming code used to analyse the data and produce the statistics. This risk is reduced through developing a good understanding of the complexities of VAT, and thoroughly reviewing and testing the programs that are used. As part of development there has been comprehensive verification exercises carried out by internal experts.

6.4 Data revision

Data revision – policy

The United Kingdom Statistics Authority (UKSA) Code of Practice for Official Statistics requires all producers of Official Statistics to publish transparent guidance on the policy for revisions.

Data revision – practice

This analysis is published annually and includes an estimate of the Scotland share of liabilities for the latest calendar year, along with revisions for the historic back series to 2011. The revisions for previous years reflect any new data received since the previous publication.

A small number of data sources used within the VA model are not available for the first publication of the estimate for that year and so are forecasted. Furthermore, some data sources used within the VA model are subject to revision. Some data sources have also been impacted by COVID-19. Therefore there is no specific point at which all the VAT liabilities for a particular year can be considered as ‘final’.

The statistics in this analysis are revised each year for the full historic series prior to the latest published year. Reasons for changes in liabilities include:

• revisions due to updating forecasted data once it is available following the first publication of an estimate

• revisions due to input data updates to historical series, for example when there are methodological improvements

Receipts statistics used for the illustrative £ million VA estimate may be revised following the end of the financial year when an annual reconciliation of receipts recorded for each tax/duty takes place ahead of publication in the HMRC Trust Statement. From this point the total receipts figure is final, subject to exceptional circumstances, such as the revisions made in June 2021.

7. Timeliness and punctuality

7.1 Timeliness

The largest data source for the Scottish VAT assignment model, the LCF Survey, is not released until 16 months after the relevant calendar year. Several other data sources for the model are also not released until around this time.

To allow time to process this data, the Scottish VAT assignment estimates cannot be published until around 19 months after the end of the calendar year.

7.2 Punctuality

In accordance with the Code of Practice for official statistics, the exact date of publication will be announced at least one calendar month before publication on both the Schedule of updates for HMRC’s statistics and the Research and statistics calendar of GOV.UK.

Any delays to the publication date will be announced on the HMRC National Statistics website.

The full publication calendar can be found on both the Schedule of updates for HMRC’s statistics and the Research and statistics calendar of GOV.UK.

7.3 Publication timeline

We plan to publish a single annual estimate in Autumn. The publication timeline is provisional and reliant on input data sources from the ONS amongst other providers, to be published to their usual timetables.

If there is a delay in the publication of these input data sources, we will need to consider if this will delay the release of our publications. This is to allow for the required level of quality assurance to release these estimates as Official Statistics.

8. Coherence and comparability

8.1 Geographical comparability

This analysis is presented for Scotland as a proportion of the United Kingdom. The United Kingdom total is calculated as a sum of regions, all calculated in the same way as Scotland.

8.2 Comparability over time

The estimates for all years included in this publication are calculated using the same methodology. Whilst methodology is adapted over time, most adaptations are applied to the back-series such that all years continue to be calculated the same way. This allows for easy and reliable comparison over time.

Some data sources or methodology changes cannot be compared over time. Examples are adjustments made for COVID-19 years.

9. Accessibility and clarity

9.1 News release

There have not been any press releases linked to this data over the past year.

9.2 Publication

The tables and associated commentary are published on the Scottish VAT assignment webpage of GOV.UK.

Tables and commentary are published in HTML which comply with the accessibility regulations set out in the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018. Further information can be found in HMRC’s accessible documents policy.

9.3 Online databases

This analysis is not used in any online databases.

9.4 Micro-data access

Access to this data is not possible in micro-data form.

9.5 Other

There aren’t any other dissemination formats available for this analysis.

9.6 Documentation on methodology

Scottish VAT assignment – Summary of VAT assignment model - GOV.UK (www.gov.uk) is publicly available to users.

9.7 Quality documentation

All official statistics produced by KAI, must meet the standards in the Code of Practice for Statistics produced by the UK Statistics Authority and all analysts adhere to best practice as set out in the ‘Quality’ pillar.

Information about quality procedures for this analysis can be found in section 4 of this document.

10. Cost and burden

HMRC and Scottish Government pay for a Scottish boost to the LCF Survey, the main survey data source used in producing the Scottish VAT assignment estimate.

11. Confidentiality

11.1 Confidentiality – policy

HMRC has a legal duty to maintain the confidentiality of taxpayer information.

Section 18(1) of the Commissioners for Revenue and Customs Act 2005 (CRCA) sets out our duty of confidentiality, which this analysis complies with.

11.2 Confidentiality – data treatment

The statistics in the publication are an estimate of the VAT raised in Scotland using various administrative/survey data, so identification of individuals is not possible.

Further information on anonymisation and data confidentiality best practice can be found on the Government Statistical Service’s website.