Official Statistics

3. Tax gaps: Excise (including alcohol, tobacco and oils)

Updated 23 June 2022

Main findings

Uncertainty is inherent in all tax gap estimates.  We assign an uncertainty rating to tax gap components on a scale using the ratings ‘very low’, ‘low’, ‘medium’, ‘high’ to ‘very high’. Details are available in the ‘Methodology and data issues’ sections of this chapter and the ‘Methodological annex’.

The Excise Duty gap is estimated at 7.2% of total theoretical tax liabilities (or £3.5 billion) in the tax year 2020 to 2021. This reflects a steady fall since the 2005 to 2006 Excise Duty gap of 8.3%.

The overall Excise tax gap on goods subject to Excise Duty and VAT is estimated at £4.5 billion (£3.5 billion in Excise Duty and £1.0 billion in VAT) in 2020 to 2021.

The overall Excise Duty gap is made up of the:

• Tobacco Duty gap, estimated at 16.2% of total theoretical tax liabilities

• Alcohol Duty gap, inclusive of beer and spirits, estimated at 9.9% of total theoretical tax liabilities

• Hydrocarbon oils Duty gap, comprised of Northern Ireland and Great Britain Diesel tax gaps, estimated at 0.5% of total theoretical tax liabilities

• Other Excise duties, including: betting and gaming, cider and perry, spirits-based ready-to-drinks and wine, estimated at 8.5% of total theoretical tax liabilities

The full data series can be seen in the online tables.

Table 3.1: Excise Duty gaps as a percentage of total theoretical tax liabilities by sector

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Tobacco	21.7%	15.6%	16.9%	14.5%	14.8%	16.6%	16.2%
Alcohol	8.0%	11.3%	9.5%	8.7%	10.2%	9.3%	9.9%
Hydrocarbon oils	2.8%	0.2%	0.5%	0.5%	0.5%	0.5%	0.5%
Other Excise duties	8.0%	9.6%	8.3%	7.1%	8.4%	7.7%	8.5%
Total	8.3%	6.2%	6.1%	5.3%	5.9%	6.0%	7.2%

Notes for Table 3.1

1. Figures for previous years have been revised.

2. Percentage figures were independently rounded to the nearest 0.1%.

3. Tobacco includes hand-rolling tobacco and cigarettes.

4. Alcohol includes beer and spirits.

5. The alcohol estimate for 2005 to 2006 is based on the beer upper bound estimate. All other years are based on the central estimate.

6. Hydrocarbon oils includes both Great Britain and Northern Ireland diesel.

7. Other Excise duties includes betting and gaming, cider and perry, spirits-based ready-to-drinks, and wine.

8. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

9. The full data series can be seen in the online tables.

Table 3.2: Excise tax gap (duty and VAT combined) by sector (£ million)

Tax gap (£ million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Tobacco	2,700	2,200	2,300	1,900	2,100	2,300	2,500
Alcohol	700	1,200	1,000	1,000	1,300	1,100	1,200
Hydrocarbon oils	830	80	180	180	180	190	150
Other Excise duties	340	730	650	560	700	630	720
Total	4,600	4,300	4,200	3,700	4,200	4,200	4,500

Notes for Table 3.2

1. Figures for previous years have been revised.

2. Figures are rounded to the nearest £100 million for tobacco and the total, to the nearest £50 million for alcohol, and to the nearest £10 million for hydrocarbon oils and Other Excise duties.

3. The tax gap includes both duty and VAT.

4. Tobacco includes hand-rolling tobacco and cigarettes.

5. Alcohol includes beer and spirits.

6. The alcohol estimate for 2005 to 2006 is based on the beer upper bound estimate. All other years are based on the central estimate.

7. Hydrocarbon oils includes both Great Britain and Northern Ireland diesel.

8. Other Excise duties includes betting and gaming, cider and perry, spirits-based ready-to-drinks, and wine.

9. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

10. The full data series can be seen in the online tables.

Alcohol

In the tax year 2020 to 2021, the Alcohol Duty gap that includes beer and spirits is 9.9% of total theoretical tax liabilities. It has varied between 7% and 13% in the time series since tax year 2005 to 2006.

The Alcohol tax gap is £1,200 million in 2020 to 2021, of which £800 million is in alcohol duties and a further £400 million in VAT.

The Alcohol Duty gap comprises both beer and spirits. The 2020 to 2021 uncertainty rating assessment for the beer tax gap estimate and the spirits tax gap estimate is ‘medium’.

Due to the uncertainty in the methodology used, the central estimates should be interpreted as an indicator of long-term trends, rather than as a precise estimate of year-on-year changes.

Users of tax gap estimates should note that the methodology used for beer differs from that of spirits in 2 key areas:

• in estimating the upper and lower bounds of the gap

• in calculating an additional uplift factor for under-reporting of consumption

Information on the different methodologies can be found in the ‘Methodological annex’.

Figure 3.1: Alcohol tax gap (duty and VAT combined) (£ million)

Notes for Figure 3.1

1. Figures for previous years have been revised.

2. Estimates for tax years 2005 to 2006 and 2006 to 2007 are based on the beer upper bound estimate. All other years are based on the central estimate.

Figure 3.1 shows the total alcohol, beer and spirits tax gaps in £ million between tax years 2005 to 2006 and 2020 to 2021. The total alcohol tax gap has increased over time from £700 million in tax year 2005 to 2006 to £1,200 million in tax year 2020 to 2021. The peak total alcohol tax gap occurred in tax year 2013 to 2014 and is estimated at £1,500 million. The beer tax gap has also increased over time from £450 million in tax year 2005 to 2006 to £1,100 million in tax year 2020 to 2021. The spirits tax gap has remained relatively flat over time, being estimated at £230 million in tax year 2005 to 2006 and £100 million in tax year 2020 to 2021. There is more volatility observed in the tax gap estimates for spirits than in the tax gap estimates for beer.

Beer

The 2020 to 2021 uncertainty rating assessment for the beer tax gap estimate is ‘medium’.

The beer illicit market share is estimated at 17% in the tax year 2020 to 2021. This results in estimated losses of £700 million in duty and a further £350 million in VAT, giving a total loss of £1,100 million (numbers do not sum due to rounding).

The illicit market share for beer has varied between 9% and 17% in the time series since tax year 2007 to 2008.

HMRC uses 2 methods to estimate the beer tax gap: an upper estimate and lower estimate. The true tax gap could be anywhere between these 2 estimates. The implied central estimate is an average of these 2 estimates and is intended to be an indicator of long-term trends.

In 2020 to 2021, the lower estimate of the beer tax gap is estimated at 8%. The lower estimate has been broadly stable between 6% and 8% between 2007 to 2008 and 2020 to 2021.

In 2020 to 2021, the upper estimate of the beer tax gap is 27%. The upper estimate has fluctuated between 9% and 27% between 2005 to 2006 and 2020 to 2021.

Table 3.3: Beer - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	9%	20%	21%	17%	24%	20%	27%
Central estimate	—	13%	14%	12%	15%	14%	17%
Lower estimate	—	7%	7%	7%	6%	7%	8%

Notes for Table 3.3

1. Figures for previous years have been revised.

2. Lower and mid-point estimates are not available for years before 2007 to 2008.

3. Lower estimates have been projected from 2016 to 2017 onwards.

4. Percentage figures were independently rounded to the nearest 1%.

5. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

6. The full data series can be seen in the online tables.

Table 3.4: Beer tax gap (£ million)

Tax gap (£million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	450	1,250	1,400	1,050	1,800	1,500	1,750
Implied central estimate	—	800	900	700	1,100	950	1,100
Implied central estimate of which VAT	—	250	250	250	350	300	350
Implied central estimate of which duty	—	550	600	500	750	700	700
Lower estimate	—	400	400	350	400	450	400

Notes for Table 3.4

1. Figures for previous years have been revised.

2. Lower and mid-point estimates are not available for years before 2007 to 2008.

3. Lower estimates have been projected from 2016 to 2017 onwards.

4. The tax gap includes both duty and VAT.

5. The pound tax gap figures were independently rounded to the nearest £50 million.

6. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

7. The full data series can be seen in the online tables.

Uncertainty rating

The uncertainty rating assessment for the beer tax gap estimate is ‘medium’. This means that the model broadly captures the tax base and its population, the methodology is reliable and uses the best available data. Assumptions in the model are closely monitored.

Areas of uncertainty, particularly in the lower bound estimate, include the model’s sensitivity to changes in the behaviours that drive under-reporting and systematic biases, with results presented as a range to represent the degree of uncertainty. Uncertainty may also arise within the upper bound model due to sampling error.

For 2020 to 2021, there is additional uncertainty from the projections of cross-border shopping and duty-free sales, beer uplift factor and lower bound estimate, but the overall rating is unchanged from ‘Measuring tax gaps 2021 edition’.

Methodology and data issues

Upper estimate

The beer tax gap upper estimate is produced using a ‘top-down’ methodology: total consumption is estimated and then legitimate consumption is subtracted, with the remainder being the illicit market.

Total consumption is estimated using the Office for National Statistics (ONS) ‘Living Costs and Food’ survey. Legitimate consumption is based on the returns that HMRC receives from the volumes of alcohol on which duties have been paid and estimates of cross-border shopping and duty-free sales.

For the tax year 2020 to 2021, estimates of cross-border shopping and duty-free sales have been projected due to the unavailability of survey data as a result of COVID-19 travel restrictions.

The uplift factor, which is used to account for under-reporting of beer consumption within the ‘Living Costs and Food’ (LCF) survey responses, has been projected in the tax year 2020 to 2021, based on an average of the previous 3 years’ uplift. This is due to data impacts caused by the COVID-19 lockdown where on-licence consumption reduced significantly and off-licence consumption increased. The uplift factor is highly sensitive to these data changes and our usual method would have resulted in a significantly higher and implausible uplift.

Lower estimate

The beer tax gap lower estimate is produced using a ‘bottom-up’ methodology. This means estimates of illicit beer are made directly, using HMRC data. This type of methodology is less comprehensive than the ‘top-down’ approach, as it does not cover all types of fraud. The methods of fraud not covered include smuggled beer, diversion of foreign-produced beer and counterfeit beer, as they are difficult to assess. This is one of the reasons why the ‘bottom-up’ approach is described as a lower-bound estimate.

The lower bound for tax years up to and including 2015 to 2016 usually only includes estimates of the diversion of UK-produced beer and drawback fraud and, as described above, is therefore an underestimate. Drawback fraud is where a refund of UK Excise Duty is obtained through deception, creating a UK tax loss.

The beer lower bound estimate has been projected from the tax year 2016 to 2017, to better reflect changes in the illicit beer market.

Implied central estimate

The implied central estimate is calculated as the average of the upper and lower estimates. Taking the mid-point of both estimates is the best approach, where we cannot measure appropriate confidence intervals. The implied central estimate is only intended as an indicator of long-term trend – the true tax gap could lie anywhere within the upper and lower bounds.

The upper and lower estimates should be interpreted as indicators of the long-term trend rather than precise estimates of the level, or year-on-year changes. They do not take account of any systematic tendency to overestimate or underestimate the tax gap that might arise from the modelling assumptions.

Revisions

Revisions to the lower bound estimate from tax years 2016 to 2017 are due to improvements in the projection method to account for operational intelligence, which better estimate changes in beer fraud. The previous projection approach was focused on diversion of beer, whereas the new approach aims to capture a wider range of beer fraud.

Spirits

The 2020 to 2021 uncertainty rating assessment for the spirits tax gap estimate is ‘medium’.

The spirits illicit market share is estimated at 2% in the tax year 2020 to 2021. This results in an estimated loss of £70 million in duty and a further £30 million in VAT, giving a total loss of £100 million.

Due to the uncertainty in the underlying data inherent in the methodology, the central estimate should be interpreted as an indicator of long-term trends, rather than a precise estimate of year-on-year changes.

Table 3.5: Spirits - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	12%	14%	9%	11%	9%	8%	8%
Central estimate	7%	7%	2%	5%	3%	2%	2%
Lower estimate	1%	0%	0%	0%	0%	0%	0%

Notes for Table 3.5

1. Figures for previous years have been revised.

2. Figures for the tax year 2020 to 2021 have been projected from 2019 to 2020.

3. Percentage figures are independently rounded to the nearest 1%.

4. Negative numbers have been truncated at zero.

5. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

6. The full data series can be seen in the online tables.

Table 3.6: Spirits tax gap (£ million)

Tax gap (£ million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	440	800	530	710	610	570	510
Central estimate	230	400	130	290	180	110	100
Central estimate of which VAT	80	150	50	110	70	40	30
Central estimate of which duty	160	250	80	180	110	70	70
Lower estimate	30	0	0	0	0	0	0

Notes for Table 3.6

1. Figures for previous years have been revised.

2. Figures for the tax year 2020 to 2021 have been projected from 2019 to 2020.

3. Negative numbers have been truncated at zero.

4. The upper, central and lower tax gap estimates contain both duty and VAT.

5. The tax gap figures are independently rounded to the nearest £10 million.

6. Figures may not appear to sum due to rounding.

7. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

8. The full data series can be seen in the online tables.

Uncertainty rating

The 2020 to 2021 uncertainty rating assessment for the spirits tax gap estimate is ‘medium’. This means that the model broadly captures the tax base and its population, the methodology is reliable (although it produces volatile results) and uses the best available data.

Areas of uncertainty include the under reporting of spirits consumption in the ONS survey data and the uplift assumption applied to counter it. Also, the central estimate may not necessarily be halfway between the upper and lower bounds, as they are confidence intervals, which may not be the same distance from the central estimate.

For 2020 to 2021 there is additional uncertainty from the projections of cross-border shopping and duty-free sales, as well as the illicit market share, however the overall rating is unchanged from the ‘Measuring tax gaps 2021 edition’.

Methodology and data issues

The spirits tax gap estimate is produced using a ‘top-down’ methodology that involves estimating total consumption, from which legitimate consumption is then subtracted, with the remainder being the illicit market.

Total consumption is estimated using data from the ONS ‘Living Costs and Food’ survey and smoothed using a centred rolling average of 3 years to reduce volatility. The spirits tax gap estimate for the tax year 2020 to 2021, however, has been projected based on the illicit market share for 2019 to 2020, as we do not have enough data to produce a 3-year rolling average. This estimate will be revised in the next edition and will be based on a rolling average of 3 years when 2021 to 2022 data is made available.

Legitimate consumption is based on the returns HMRC receives from the volumes of alcohol on which duties have been paid and an estimate of cross-border and duty-free shopping. For the tax year 2020 to 2021, estimates of cross-border shopping and duty-free sales have been projected due to the unavailability of survey data, as a result of COVID-19 travel restrictions.

The central estimate should be interpreted as an indicator of the long-term trend in the illicit market share, rather than a precise estimate of the level, or year-on-year changes. The confidence intervals indicate the uncertainty surrounding the estimate due to sampling error. They do not account for additional types of error that may arise from the assumptions made in the calculations. Corrections have been made to the calculation of the confidence intervals across the times series.

Revisions

Revisions to the spirits tax gap for the tax year 2019 to 2020 are due to new available data in 2020 to 2021 that has allowed us to apply a 3-year rolling average. This replaces the approach used in the ‘Measuring tax gaps 2021 edition’ where we applied a 2-year rolling average. As discussed above, we are now projecting the latest year based on the illicit market share of the previous year instead of a 2-year moving average.

Revisions have also been made to the upper and lower estimates to account for data that was missed from quarter 1 of the LCF Survey across the time series.

Wine

It has not been possible to estimate the illicit market share for wine, due to the unavailability of a data source previously used to estimate the wine tax gap. We have included wine within our illustrative tax gap estimate for ‘Other Excise duties’.

The wine tax gap was previously estimated using a ‘top-down’ methodology. As with the spirits estimate and the beer upper-bound estimate, legitimate consumption was subtracted from estimated total consumption, with the remainder being the illicit market.

There are additional considerations when estimating total wine consumption compared to spirits and beer. HMRC can use the LCF survey to estimate consumption of beer and spirits, however, the LCF is a survey of households and there is a sizeable non-household component of spending on wine. For example, businesses will spend money on wine to entertain clients and potential clients. This business expenditure is not captured by the LCF survey of households.

Furthermore, although the LCF is a representative survey of households, it may not adequately capture infrequent, but high-volume, expenditure on wine; for example, the cost of providing wine at a wedding or other large functions. HMRC previously used commercial data to estimate these additional components of wine expenditure. However, due to a change in the underlying data source, we cannot use the existing methodology to estimate the wine tax gap from tax years 2013 to 2014 onwards.

HMRC produced an illustrative estimate of the wine tax gap, from 2013 to 2014 and 2016 to 2017 inclusive, by taking the average illicit market share from tax years 2011 to 2012 and 2012 to 2013. However, this method produced a static estimate of the tax gap, when the wider alcohol market and tax gap has changed. To reflect this change, the wine tax gap is now included in the estimate for ‘Other Excise’. Further details of the method are in the ‘Other Excise duties’ section.

Tobacco

The 2020 to 2021 uncertainty assessment for cigarettes and hand-rolling tobacco tax gap estimate is ‘high’.

In the tax year 2020 to 2021, the tobacco duty tax gap that comprises cigarettes and hand-rolling tobacco is estimated at 16% of total theoretical tax liabilities. It has varied between 12% and 22% in the time series since tax year 2005 to 2006.

The tobacco tax gap is estimated to be £2.5 billion in 2020 to 2021. An estimated total of £1.9 billion has been lost in tobacco duties and a further £0.5 billion in VAT (numbers do not sum due to rounding).

Due to the uncertainty in the underlying data inherent in the methodology, the central estimate should be interpreted as an indicator of long-term trends rather than a precise estimate of year-on-year changes.

Results since 2018 to 2019 are projected based on the tax year 2017 to 2018 tobacco tax gap percentage. Our model that produces this estimate uses the ONS ‘Opinions and Lifestyle’ survey to estimate total consumption. There is a break in the consistency of the data, as both the mode of sampling, and questions about consumption have changed. For this reason, further development is required to understand the implications to our modelling. For more details on changes to the survey, see the ‘Methodological annex’.

The combined tobacco tax gap is the sum of the central estimates for cigarettes and hand-rolling tobacco. It is not possible to combine both upper (or equivalently, lower) estimates, so there is no upper or lower bound estimate for the combined tobacco tax gap. More detail on upper and lower bounds is given in the ‘Methodology and data issues’ section and in the ‘Methodological annex’.

Table 3.7: Tobacco tax gap - Combined (£ million)

Combined tobacco illicit market	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Combined tobacco illicit market	2,700	2,200	2,300	1,900	2,100	2,300	2,500

Notes for Table 3.7

1. Figures are independently rounded to the nearest £100 million.

2. Figures include both duty and VAT.

3. The upper (or lower) bounds for both cigarette and hand-rolling tobacco cannot be combined. Only the central estimates are simultaneously possible.

4. Figures since 2018 to 2019 are projected based on the 2017 to 2018 cigarette and hand-rolling tobacco tax gap percentages.

5. Numbers may not sum due to rounding.

6. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

7. The full data series can be seen in the online tables.

Table 3.8: Tobacco tax gap - Cigarettes (£ million)

Cigarette illicit market	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	2,400	2,100	2,100	1,300	1,400	1,400	1,200
Central estimate	1,900	1,500	1,600	900	1,000	1,000	900
Central estimate of which VAT	300	300	300	200	200	200	200
Central estimate of which duty	1,500	1,200	1,300	800	800	800	700
Lower estimate	1,400	900	1,100	600	600	600	500

Notes for Table 3.8

1. Figures are independently rounded to the nearest £100 million.

2. Figures for the upper, central and lower estimates include both duty and VAT.

3. The cigarettes illicit market figures are based on weighted average price (WAP) data for all UK duty paid cigarettes.

4. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

5. Numbers may not sum due to rounding.

6. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

7. The full data series can be seen in the online tables.

Table 3.9: Tobacco tax gap - Hand-rolling tobacco (£ million)

Hand-rolling tobacco illicit market	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	1,000	900	800	1,100	1,200	1,400	1,800
Central estimate	800	700	700	1,000	1,100	1,300	1,600
Central estimate of which VAT	200	200	200	200	300	300	400
Central estimate of which duty	700	600	500	700	800	900	1,200
Lower estimate	700	500	500	800	900	1,100	1,400

Notes for Table 3.9

1. Figures are independently rounded to the nearest £100 million.

2. Figures for the upper, central and lower estimates include both duty and VAT.

3. The weighted average price (WAP) data for all UK duty paid hand-rolling tobacco is not available prior to tax year 2012 to 2013, so the losses are based on the price of a ‘typical brand’.

4. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

5. Numbers may not sum due to rounding.

6. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

7. The full data series can be seen in the online tables.

Cigarettes

The cigarette illicit market share is estimated at 9% in the tax year 2020 to 2021, which is projected based on the tax year 2017 to 2018 tax gap percentage. This resulted in an estimated loss of £700 million in duty and a further £200 million in VAT, giving a total loss of £900 million.

Total consumption is obtained from the ‘Opinions and Lifestyle’ survey conducted by the ONS. The sample is representative of the cigarette-smoking population, but as with all surveys, there is sampling error. Due to the nature of the data, small movements in overall consumption can have noticeable impacts on the illicit market and tax gap.

UK tax-paid cigarette consumption has declined steadily from 49.5 billion cigarettes in tax year 2005 to 2006 to 23.5 billion cigarettes in 2020 to 2021. There has been a long-term decline in the estimated volume of cigarettes in the illicit market since 2005 to 2006 when the central estimate for the illicit market was 10.0 billion cigarettes in the UK. The central illicit market estimate has been fairly stable since the tax year 2014 to 2015, ranging between 2.5 billion and 5.0 billion cigarettes. The central estimate should be interpreted in the context of statistical uncertainty.

Figure 3.2: Cigarette central estimate for illicit market and UK tax-paid consumption (billion cigarettes)

Figure 3.2 shows the cigarette UK tax paid consumption and illicit market in billions of cigarettes between tax years 2005 to 2006 and 2020 to 2021. The UK tax-paid consumption has decreased from 49.5 billion cigarettes in tax year 2005 to 2006 to 23.5 billion cigarettes in tax year 2020 to 2021. The illicit market has decreased from 10.0 billion cigarettes in tax year 2005 to 2006 to 2.5 billion cigarettes in tax year 2020 to 2021, and has been fairly stable since 2014 to 2015.

Table 3.10: Cigarettes - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	19%	16%	17%	12%	12%	12%	12%
Central estimate	16%	12%	14%	9%	9%	9%	9%
Lower estimate	12%	8%	10%	6%	6%	6%	6%

Notes for Table 3.10

1. Figures are independently rounded to the nearest 1%.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table 3.11: Cigarettes - total consumption volume (billion cigarettes)

Total consumption	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	66.5	40.5	37.5	32.5	31.5	31.0	27.0
Central estimate	64.0	38.5	36.0	31.0	30.0	30.0	26.0
Lower estimate	61.5	36.5	34.5	30.0	29.0	29.0	25.0
UK tax paid consumption	49.5	32.0	29.5	27.0	26.5	26.0	23.5

Notes for Table 3.11

1. Figures are rounded to the nearest 0.5 billion cigarettes.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table 3.12: Cigarettes - illicit market volume (billion cigarettes)

Illicit market	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	12.5	6.5	6.5	4.0	4.0	3.5	3.0
Central estimate	10.0	4.5	5.0	3.0	2.5	2.5	2.5
Lower estimate	7.5	2.5	3.5	1.5	1.5	1.5	1.5

Notes for Table 3.12

1. Figures are independently rounded to the nearest 0.5 billion cigarettes.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table 3.13: Cigarettes - cross-border shopping volume (billion cigarettes)

Cross-border shopping	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Cross-border shopping	4.0	1.5	1.5	1.5	1.0	1.5	<0.5

Notes for Table 3.13

1. Figures are independently rounded to the nearest 0.5 billion cigarettes.

2. Cross-border shopping includes duty-free as well as EU duty paid.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Hand-rolling tobacco

The hand-rolling tobacco illicit market share is estimated at 34% in the tax year 2020 to 2021. This resulted in an estimated loss of £1,200 million in duty and a further £400 million in VAT, giving a total loss of £1,600 million.

The UK tax-paid consumption volume of hand-rolling tobacco increased steadily between tax years 2005 to 2006 and 2013 to 2014 after which it was relatively stable, with further increases in 2019 to 2020 and 2020 to 2021. The illicit market volume of hand-rolling tobacco has shown a long-term decline since 2005 to 2006. It has increased since 2016 to 2017 and is estimated to be 4.6 million kg in 2020 to 2021.

Figure 3.3: Hand-rolling tobacco central estimate for illicit market and UK tax paid consumption (million kg)

Figure 3.3 shows the UK tax paid consumption and illicit market of hand-rolling tobacco in million kg between tax years 2005 to 2006 and 2020 to 2021. The UK tax paid consumption has increased from 3.2 million kg in tax year 2005 to 2006 to 8.7 million kg in tax year 2020 to 2021. The illicit market has decreased from 6.1 million kg in tax year 2005 to 2006 to 4.6 million kg in tax year 2020 to 2021.

Table 3.14: Hand-rolling tobacco - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	65%	36%	31%	37%	37%	37%	37%
Central estimate	60%	30%	26%	34%	34%	34%	34%
Lower estimate	55%	25%	21%	31%	31%	31%	31%

Notes for Table 3.14

1. Figures are independently rounded to the nearest 1%.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table 3.15: Hand-rolling tobacco - total consumption volume (million kg)

Total consumption	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	11.2	10.5	10.2	10.4	11.0	12.2	14.0
Central estimate	10.2	9.8	9.5	10.0	10.6	11.7	13.4
Lower estimate	9.1	9.0	8.9	9.6	10.1	11.2	12.8
UK tax paid consumption	3.2	6.2	6.4	6.1	6.4	7.2	8.7

Notes for Table 3.15

1. Figures are independently rounded to the nearest 100,000 kg.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table 3.16: Hand-rolling tobacco - illicit market volume (million kg)

Illicit market	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	7.2	3.8	3.2	3.9	4.1	4.5	5.2
Central estimate	6.1	3.0	2.5	3.4	3.6	4.0	4.6
Lower estimate	5.0	2.2	1.9	3.0	3.2	3.5	4.0

Notes for Table 3.16

1. Figures are independently rounded to the nearest 100,000 kg.

2. Figures since 2018 to 2019 are projected based on the 2017 to 2018 tax gap percentage.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Table: 3.17: Hand-rolling tobacco - cross-border shopping volume (million kg)

Cross-border shopping	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Associated with upper estimate	0.9	0.6	0.6	0.5	0.5	0.5	0.1
Associated with central estimate	0.8	0.6	0.6	0.4	0.5	0.5	0.1
Associated with lower estimate	0.7	0.6	0.6	0.4	0.5	0.5	0.1

Notes for Table 3.17

1. Figures are independently rounded to the nearest 100,000 kg.

2. Cross-border shopping includes duty-free as well as EU duty paid.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Uncertainty rating

The 2020 to 2021 uncertainty assessment for cigarettes and hand-rolling tobacco tax gap estimate is ‘high’. This means that the model captures the majority of the tax base and its population, the model is satisfactory and uses ONS consumption data that requires high risk assumptions to be applied for it to be suitable.

Areas of uncertainty include the high-risk consumption uplifts that account for under-reporting and excluded cigarette consumption, and projections applied due to changes in the ONS consumption data. The upper and lower estimates contain random variations.

Since ‘Measuring tax gaps 2021 edition’, the rating has been increased to ‘high’ for ‘Measuring tax gaps 2022 edition’ as the estimates are now deemed illustrative, given that they are being projected for an additional year.

Methodology and data issues

The estimates are produced using a ‘top-down’ methodology: total consumption is estimated, legitimate consumption is subtracted, and the remainder is the illicit market.

Total tobacco consumption is estimated using data from the ONS ‘Opinions and Lifestyle’ survey. As with all surveys, it contains levels of sampling error and limitations in accurately recording self-reported behaviour, which can lead to uncertainty in results. In 2018, methodological changes were made to the ‘Opinions and Lifestyle’ survey, which caused a discontinuity in the time series. Consequently, results since 2018 to 2019 are projected based on the tax year 2017 to 2018 tobacco tax gap percentage.

Legitimate tobacco consumption is based on the returns that HMRC receives from the volumes of tobacco on which duties have been paid and an estimate of cross-border shopping and duty-free sales. For tax year 2020 to 2021, estimates of cross-border shopping and duty-free sales have been projected due to unavailability of survey data as a result of COVID-19 travel restrictions.

In addition to the uplift that accounts for under-reporting in tobacco consumption, there is an uplift that accounts for people who falsely deny smoking, which comes from the ‘Health Survey for England’.

The estimates are presented within a range defined by the upper and lower estimates. The range provides an indication of the size of the illicit markets in cigarettes and hand-rolling tobacco, depending on how dual smoker survey respondents (someone who smokes both cigarettes and hand-rolling tobacco) are treated.

The upper bound for cigarettes corresponds to a scenario where most dual smokers consume cigarettes, whereas the upper bound for hand-rolling tobacco corresponds to a scenario where most dual smokers consume hand-rolling tobacco. The lower bounds for cigarettes and hand-rolling tobacco are calculated in a similar way. Given the differing scenarios applied to cigarettes and hand-rolling tobacco, it is not possible to append these to form an upper and lower bound for overall tobacco estimates.

Oils

The oils tax gap consists of Great Britain (GB) and Northern Ireland (NI) diesel, but excludes petrol, which is assumed to be negligible due to low demand from commercial sectors and the flammable nature of the product.

The 2020 to 2021 uncertainty rating for the diesel tax gap estimate is ‘low’.

The diesel illicit market share is estimated at 1.0% in tax year 2020 to 2021. This resulted in an estimated loss of £120 million in duty and a further £30 million in VAT, giving a total loss of £150 million.

A new methodology was introduced to estimate the diesel tax gap from tax year 2016 to 2017, therefore the figures from 2016 to 2017 are not directly comparable to previous years. More details on the methodology are given in the ‘Methodological annex’.

Great Britain diesel

The GB diesel illicit market share is estimated at 1% in the tax year 2020 to 2021. This resulted in an estimated loss of £100 million in duty and a further £30 million in VAT, giving a total loss of £130 million.

Table 3.18: Great Britain diesel - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	8%	3%	1%	1%	1%	1%	1%
Central estimate	5%	1%	1%	1%	1%	1%	1%
Lower estimate	2%	0%	<1%	<1%	<1%	<1%	<1%

Notes for Table 3.18

1. Figures for previous years have been revised.

2. Figures on the illicit market share are independently rounded to the nearest 1%.

3. Negative numbers have been truncated at zero for the lower estimate.

4. Figures for the tax year 2015 to 2016 are rolled forward from 2013 to 2014.

5. Figures from 2016 to 2017 are not directly comparable to previous years due to a methodological change.

6. The same rate of non-compliance in 2016 to 2017 is applied to diesel receipts in tax years 2017 to 2018 and 2018 to 2019. The same rate of non-compliance in 2019 to 2020 is applied to diesel receipts in the tax year 2020 to 2021.

7. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

8. The full data series can be seen in the online tables.

Table 3.19: Great Britain diesel - tax gap (£ million)

Tax gap (£ million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	1,250	700	230	230	230	250	200
Central estimate	750	<50	140	140	140	160	130
Central estimate of which VAT	200	<50	30	30	30	40	30
Central estimate of which duty	600	<50	110	110	110	120	100
Lower estimate	300	0	60	60	60	70	50

Notes for Table 3.19

1. Figures for previous years have been revised.

2. The upper, central and lower tax gap estimates include both duty and VAT.

3. The figures are rounded to the nearest £10 million for years following 2016 to 2017 and £50 million for years prior.

4. Negative numbers have been truncated at zero for the lower estimate.

5. Figures for the tax year 2015 to 2016 are rolled forward from 2013 to 2014.

6. Figures from 2016 to 2017 are not directly comparable to previous years, due to a methodological change.

7. The same rate of non-compliance in 2016 to 2017 is applied to diesel receipts in tax years 2017 to 2018 and 2018 to 2019. The same rate of non-compliance in 2019 to 2020 is applied to diesel receipts in tax year 2020 to 2021.

8. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

9. The full data series can be seen in the online tables.

Northern Ireland diesel

The NI diesel illicit market share is estimated at 4% in tax year 2020 to 2021. This resulted in an estimated loss of £20 million in duty and a further £10 million in VAT, giving a total loss of £20 million (numbers do not sum due to rounding).

Table 3.20: Northern Ireland diesel - illicit market share

Illicit market share	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	21%	11%	7%	7%	6%	5%	6%
Central estimate	19%	8%	5%	5%	5%	4%	4%
Lower estimate	16%	4%	3%	3%	3%	2%	2%

Notes for Table 3.20

1. Figures for previous years have been revised.

2. Figures for the illicit market share are independently rounded to the nearest 1%.

3. Figures for the tax year 2015 to 2016 are rolled forward from 2013 to 2014.

4. Figures from tax year 2016 to 2017 are not directly comparable to previous years due to a methodological change.

5. The same rate of non-compliance in 2016 to 2017 is applied to diesel receipts in tax years 2017 to 2018 and 2018 to 2019. The same rate of non-compliance in 2019 to 2020 is applied to diesel receipts in the tax year 2020 to 2021.

6. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

7. The full data series can be seen in the online tables.

Table 3.21: Northern Ireland diesel - tax gap (£ million)

Tax gap (£ million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Upper estimate	100	70	60	60	60	50	40
Central estimate	80	50	40	40	40	30	20
Central estimate of which VAT	20	10	10	10	10	10	10
Central estimate of which duty	70	30	30	30	30	20	20
Lower estimate	70	20	20	20	20	20	10

Notes for Table 3.21

1. Figures for previous years have been revised.

2. Figures for the illicit market share are independently rounded to the nearest £10 million.

3. The upper, central and lower tax gap estimates includes both duty and VAT.

4. Figures for the tax year 2015 to 2016 are rolled forward from 2013 to 2014.

5. Figures from tax year 2016 to 2017 are not directly comparable to previous years due to a methodological change.

6. The same rate of non-compliance in 2016 to 2017 is applied to diesel receipts in tax years 2017 to 2018 and 2018 to 2019. The same rate of non-compliance in 2019 to 2020 is applied to diesel receipts in tax year 2020 to 2021.

7. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

8. The full data series can be seen in the online tables.

Uncertainty rating

The 2020 to 2021 uncertainty rating for the diesel tax gap estimate is ‘low’. This means that the model captures the majority of the tax base and its population, the methodology is robust, and data are reliable and suitable for purpose.

Areas of uncertainty are the assumptions used to project the estimate that account for the 3-year random enquiry survey frequency.

There is some additional uncertainty for 2020 to 2021 as the same rate of non-compliance is used for an additional year, but the overall rating is unchanged from ‘Measuring tax gaps 2021 edition’.

Methodology and data issues

A ‘bottom-up’ methodology to estimate the diesel tax gap was introduced from tax year 2016 to 2017, therefore the figures from 2016 to 2017 are not directly comparable to previous years.

Legitimate consumption is based on the returns that HMRC receives from the volumes of diesel on which duties have been paid. Illicit consumption is estimated using the proportion of vehicles found to be misusing rebated fuel in a random sample survey conducted by HMRC in 2017 and 2020. Revenue losses (gross tax gap) associated with illicit consumption are estimated using average retail prices, duty rates and VAT rates. The net tax gap is then calculated as the gross tax gap minus compliance yield.

The upper and lower estimates correspond to confidence intervals that indicate the range where the true value of the illicit market may lie and arises due to random sampling errors.

Revisions

The oils duty receipts number for 2019 to 2020 was incorrectly rounded in the ‘Measuring tax gaps 2021 edition’, but has now been replaced with the actual unrounded number, leading to a small revision in the tax gap £ estimate for 2019 to 2020.

Other Excise duties

The Other Excise duties tax gap, which includes betting and gaming, cider and perry, spirits-based ready-to-drink beverages and wine duties, is based on an experimental method.

The 2020 to 2021 uncertainty rating for Other Excise duties tax gap estimate is ‘very high’.

The tax gap for Other Excise duties in the tax year 2020 to 2021 is estimated at 8% of total theoretical tax liabilities. This resulted in an estimated loss of £720 million.

Table 3.22: Other Excise duties tax gap (£ million)

Tax gap (£ million)	Baseline 2005-06	2015-16	2016-17	2017-18	2018-19	2019-20	2020-21
Other Excise duties	340	730	650	560	700	630	720

Notes for Table 3.22

1. Figures for previous years have been revised.

2. Figures rounded to nearest £10 million.

3. The baseline year 2005 to 2006 has been included to illustrate the long-term trend.

4. The full data series can be seen in the online tables.

Uncertainty rating

The 2020 to 2021 uncertainty rating for Other Excise duties tax gap estimate is ‘very high’. This means that the model may not capture the appropriate tax base and population, the experimental methodology is heavily assumption-based and data may not be representative.

Areas of uncertainty include the omission of non-compliance specific to the Other Excise duty population and the use of a simple methodology that uses receipts data and tax gap estimates from other excise tax gap models.

The rating is unchanged from ‘Measuring tax gaps 2021 edition’.

It is not possible to produce a range of estimates for Other Excise duties due to the experimental methodology applied.

Methodology and data issues

A proxy indicator for the scale of revenue losses across Other Excise duties has been produced by estimating the weighted average percentage of tax gaps for similar taxes each year and multiplying this by total theoretical tax liabilities (receipts plus tax gap).

The average percentage revenue losses should not be considered estimates of the true percentage loss across betting and gaming, cider and perry, spirits-based ready-to-drink beverages and wine, as this is unknown. Many of these taxes are very different from one another in nature and are, therefore, subject to different rules. The true percentage tax gaps are therefore likely to vary across the taxes.

Revisions

Small revisions have been made due to changes in the underlying alcohol models that feed into the assumptions.

Go directly to other pages in this report:

• Tax gaps: Main findings
• 1. Tax gaps: Summary
• 2. Tax gaps: VAT
• 4. Tax gaps: Income Tax, National Insurance contributions and Capital Gains Tax
• 5. Tax gaps: Corporation Tax
• 6. Tax gaps: Other taxes
• 7. Tax gaps: Illustrative tax gap by behaviour

Complete the HMRC Measuring tax gaps 2022 user survey.