Taxes on different forms of consumer spending provide the second biggest source of revenue for government, with VAT (value added tax) by far the biggest of those. In our latest forecast, we expect VAT to raise £136.6 billion in 2019-20 (this measure of VAT excludes refunds of VAT made to certain public sector organisations). That would represent 16.8 per cent of all receipts and is equivalent to around £4,800 per household and 6.2 per cent of national income.

VAT is levied on the purchase of many goods and services. It is reflected in the price paid when items are bought and is collected from traders. Unlike a simple sales tax, it is actually levied – as the name implies – on the amount of value added at each stage of the production chain. For example, a retailer would not be liable for the full amount of VAT if the good they sold had been purchased from a wholesaler.

The standard rate of VAT is 20 per cent, with around half of household expenditure subject to this rate. The reduced rate is 5 per cent and is applied to domestic fuel and power, children’s car seats and some other goods. Around 3 per cent of expenditure is taxed at this reduced rate. Other goods and services, such as books, newspapers, children’s clothing and many foods attract no VAT (i.e. they are either exempt from VAT or are ‘zero-rated’).

  • Latest forecast

    We expect VAT receipts to rise steadily in cash terms and to be relatively flat as a share of GDP. With relatively few policy measures affecting the forecast, that is driven almost entirely by our forecast for nominal consumer spending, which we expect to make up a relatively stable share of GDP over the forecast period.

    More detail on our latest forecast and how it was revised relative to our previous forecast in November was provided in paragraphs 4.32 to 4.34 of our March 2019 EFO.

    Expand to read the extract from our March 2019 EFO

    Back to top

  • Latest monthly data

    VAT receipts are spread fairly evenly throughout the year, with the slight peak in December reflecting the Christmas shopping season.

    Back to top

  • Forecast methodology

    Forecast process

    The OBR commissions forecasts of VAT receipts from HM Revenue and Customs for each fiscal event. The forecasts start by generating an in-year estimate for receipts in the current year, then uses a model to forecast growth in receipts from that starting point. We provide HMRC with economic forecasts that are then used to generate the tax forecasts. These are scrutinised in a challenge process that typically involves two rounds of meetings where HMRC analysts present forecasts to the Budget Responsibility Committee and OBR staff. This process allows the BRC to refine the assumptions and judgements that underpin the forecasts before they are published in our Economic and fiscal outlooks.

    Forecasting models

    The VAT forecast is based on the concept of a ‘VAT theoretical liability’ or VTTL – the total value of VAT that could theoretically be collected from the tax base. For outturn years, this is calculated using ONS data on total expenditure by sector. Respective VAT rates are applied to detailed breakdowns of this expenditure, to calculate a total theoretical liability. This liability is split into four main sectors: household, exempt, government and household investment. Each sector is then grown in line with the relevant elements of our economy and fiscal forecasts.

    A key factor in forecasting VAT receipts is the standard-rated share (SRS) – the proportion of consumer spending taxed at the standard 20 per cent rate. The latest outturn is obtained from expenditure data. Changes in the SRS are then projected using an econometric model, primarily driven by our forecast for spending on durable goods (such as cars).

    The difference between the VTTL and actual receipts in the last outturn year represents an implied ‘VAT gap’. The gap is made up of error, fraud, evasion, avoidance and debts owed by firms to HMRC, as well as any errors in estimating the VTTL itself. After taking a judgement on the VAT gap next year, we typically assume this gap remains constant as a proportion of the VTTL over the forecast. We then make further adjustments to reflect new or yet-to-be implemented compliance measures over the forecast. Combining the VTTL forecast and the VAT gap projection gives the final VAT forecast.

    Main forecast determinants

    The main determinants of our VAT forecast are those related to the tax base (in particular consumer spending). See the ready reckoners section below for more information on the effects of these determinants on income tax receipts

    Main forecast judgements

    The most important judgements in our VAT forecast are related to the economy forecast – particularly nominal consumer spending. Alongside, we need to make several other judgements. These include:

    • In-year estimate – year-to-date receipts provide valuable information about the likely outturn for the current year. Judgements are made whether to push any unexplained shortfall or overshoot to future years;
    • Implied VAT gap – in general held flat prior to incorporating compliance measures over the forecast period. We could re-assess this assumption in light of information on compliance risks and trends in VAT debt; and
    • Standard-rated share –the share of consumer spending subject to the standard rate of VAT is a key assumption. Our judgement is informed by an econometric forecast model, the latest historical estimates of the share and any other factors in our broader economy forecast that we consider relevant.

    Back to top

  • Previous forecasts

    VAT receipts have typically been stronger than our forecasts in recent years. Nominal household consumption growth was weaker than we expected in our earlier forecasts, but has been stronger more recently. Abstracting from this (and from other elements in the tax base) the composition of household spending has played a small role as the share of spending subject to the standard rate has held up. This strength also reflects the implied VAT gap closing by more than we had assumed in our earlier forecasts.

    Prior to November 2015, there was an error in the VAT deductions element of the model which had overstated the scale of deductions to the government sector and hence understated total VAT receipts. This was identified in our 2015 Forecast evaluation report and corrected in the subsequent Economic and fiscal outlook. Forecast differences since November 2015 have been significantly lower.

    Back to top

  • Policy measures

    Since our first forecast in June 2010, governments have announced 45 policy measures affecting our forecast for VAT. The original costings for these measures are contained in our policy measures database and were described briefly in the Treasury’s relevant Policy costings document. For measures announced since December 2014, the uncertainty ranking that we assigned to each is set out in a separate database. For those deemed ‘high’ or ‘very high’ uncertainty, the rationale for that ranking was set out in Annex A of the relevant Economic and fiscal outlook.

    Key VAT policy changes since 2010 have included:

    • The rise in the standard rate of VAT to 20 per cent from January 2011 announced in the June 2010 Budget. This was expected to raise VAT receipts by at least £12 billion from 2011-12 onwards;
    • Anti-avoidance measures – These include measures aimed at closing loopholes, reducing the level of tax avoidance and evasion, and enhancing the compliance performance of HMRC. For example, measures announced in the July 2015 Budget and the 2015 Autumn Statement are expected to raise around £1¾ billion by 2020-21; and
    • Making Tax Digital – This is an HMRC initiative to interact digitally with small businesses across taxes, working with the private sector to introduce software that will design out record-keeping errors in taxpayers’ returns. Use of making tax digital is mandated for VAT from 2019-20 onwards for companies above the VAT registration threshold.

    Back to top

  • Ready reckoners

    ‘Ready reckoners’ show how our fiscal forecasts could be affected by changes in selected economic determinants. They are stylised quantifications that reflect the typical impact of changes in economic variables on receipts and spending. These estimates are specific to our March 2017 forecast and we would expect them to become outdated over time, as the economy and public finances, and the policy setting, continue to evolve. They are subject to uncertainty because they are based on models that draw on historical relationships or simulations of policy settings. More information can be found in the ‘Tax and spending ready reckoners’ spreadsheet we published alongside our 2017 Fiscal risks report.

    The table below shows that:

    • around 70 per cent of VAT receipts are derived from household consumption, and these move largely one-for-one with changes in nominal consumer spending;
    • changes in the composition of consumption are also important, as different types of spending attract different VAT rates. The ‘standard rated share’ (SRS) is the share of nominal household consumer spending subject to the standard rate of VAT; and
    • changes in spending from other sectors generally have smaller impacts on receipts. These sectors include the exempt, government and household investment sectors. Changes in nominal GDP, government consumption and household investment growth will feed into these sectors respectively.

    Back to top

  • Other information

    Model review

    In our 2017 Forecast evaluation report we set out the initial conclusions of our review of fiscal forecasting models. This section sets out the conclusions of our review for the main VAT models (the ‘theoretical tax liability’ (VTTL) model and the ‘standard rated share’ (SRS) model).

    VAT is levied on the purchase of most goods and services. It is reflected in the price paid when items are bought and is collected from traders. The VAT model is based on the concept of a ‘VAT theoretical tax liability’ or VTTL – the total value of VAT that could theoretically be collected from the tax base. Each sector within this tax base is then grown in line with the relevant elements of our economy and fiscal forecasts. The main forecasting challenge relates to the standard-rated share (SRS) of each sector of the tax base – the proportion taxed at the standard rate of 20 per cent. In our March 2017 forecast, we overhauled our household-sector SRS model to better capture recent trends and to ensure better consistency with our overall economy forecast.

    The average absolute two-year ahead fiscal forecasting difference since our March 2012 forecast has been 2.4 per cent, which on a volatility adjusted basis is lower than the average across our fiscal forecasts. The average difference has been positive at 2.4 per cent, smaller than average, with all the past four March forecasts under-predicting receipts at the two-year horizon. A significant proportion of these under-predictions in recent years reflect an error related to the projection of government VAT deductions that we identified in our 2015 FER and corrected in our November 2015 forecast.

    We have agreed the following priority with HMRC for development work over the coming year:

    • Review the data used in the historical theoretical tax liability model to inform our forecast judgements.

    We have worked with HMRC to better understand the implications of the recent Blue Book changes to the modelling of the VAT theoretical tax liability (VTTL) and their implications for the VAT gap. Our next priority is to review the non-household sectors of the VTTL model.

    Back to top

  • Boxset