New modelling of alcohol pricing policies, alcohol consumption and harm in Wales: summary

This report was commissioned in July 2024 by the Welsh Government to build on previous analyses, published in 2018, that modelled the impact of the initial introduction of a Minimum Price for Alcohol (MPA) in Wales.

The aims of this new analysis were to use the latest available data and the most recent version of the Sheffield Tobacco and Alcohol Policy Model to assess alternative options for the future of MPA and to answer the following research questions.

What is the estimated impact of increasing the current 50p/unit MPA threshold to thresholds ranging from 55p-80p/unit, reducing the MPA threshold to 40p or 45p/unit, or removing MPA entirely?

What changes in alcohol duties would be required to achieve the same impact on alcohol-specific deaths in the overall population, or in the most deprived quintile, as each of the alternative MPA thresholds modelled?

What would the future impact be of alternative options to uprating (or not) the MPA threshold beyond 2025 under alternative mechanisms for this uprating, assuming, for illustrative purposes, that the MPA threshold were raised to 65p/unit in 2026.

Baseline alcohol consumption, purchasing and harm

18% of adults in Wales do not drink alcohol. Of those that do drink, 77% drink at moderate levels – within the UK low risk guidelines - 19% are hazardous drinkers and 4% drink at harmful levels ^{[footnote 1]}. The 23% of adult drinkers who drink above the guideline levels consume 70% of all alcohol, with the heaviest drinking 4% accounting for 27% of all alcohol drunk in Wales.

Moderate drinkers consume an average of 4.5 units and spend £7.62 on alcohol per week. This compares to an average of 25 units and £32.20 per week for hazardous drinkers and 72 units and £77.68 for harmful drinkers.

Those in the most deprived quintile of the Welsh population are less likely to drink than those in the least deprived, consume less on average (11.5 units per week vs 12.2) and spend less on alcohol (£15.06 vs £16.89).

Hazardous and harmful drinkers pay substantially less, on average, for alcohol than moderate drinkers (£1.29 and £1.08 respectively compared to £1.71 per unit). This is through a combination of consuming a smaller proportion of their alcohol in the on-trade, where prices are generally higher, buying a greater proportion of their alcohol as beer and cider, which are generally cheaper per unit, and buying cheaper products within the same category.

Drinkers from lower socioeconomic groups pay less, on average, for alcohol compared to drinkers from higher socioeconomic groups, but these differences are much smaller than between moderate and hazardous/harmful drinkers. For example, the mean price paid for people from the most deprived quintile is £1.31 per unit compared to £1.38 for the least deprived quintile.

Every year an estimated 699 people die as a direct consequence of their alcohol consumption, the equivalent of 23,550 years of life lost. There are also an estimated 11,243 hospital admissions attributable to alcohol annually.

These harms are heavily concentrated in the heaviest drinkers, with the 4.3% of drinkers drinking at harmful levels accounting for 30% of alcohol-attributable admissions, 48% of alcohol-attributable deaths, and 42% of years of life lost to alcohol.

Alcohol harms are also substantially higher in the most deprived groups, despite them consuming less alcohol, on average. Overall rates of alcohol-attributable deaths in the most deprived quintile of the population are 3.1 times higher than in the least deprived quintile and rates of alcohol-attributable hospital admissions are 1.3 times higher.

Modelled effects of changes to the MPA threshold

Increasing the MPA threshold to 65p per unit would reduce population level alcohol consumption by an estimated 0.3 units per drinker per week, a 2.7% reduction. This reduction would be smaller for lower MPA thresholds (e.g. -0.6% for 55p) and larger for higher thresholds (up to -9.4% for an 80p MPA). Reducing the MPA threshold is estimated to increase alcohol consumption (e.g. +0.3% for 40p) and removing MPA entirely is estimated to lead to larger increases (+0.7%).

Increasing the MPA threshold is estimated to reduce the number of people drinking at hazardous and harmful levels (e.g. a 5.7% reduction in the number of harmful drinkers for a 65p MPA), while reducing or removing it would have the opposite effect (e.g. a 1.1% increase in harmful drinkers if MPA was removed entirely).

For all modelled MPA thresholds, the largest changes in alcohol consumption are estimated to occur in the most deprived groups (e.g. -4.0% for the most deprived quintile vs -1.8% for the least deprived under a 65p MPA, or +1.3% and +0.4% respectively if MPA was removed entirely).

Overall consumer spending on alcohol is estimated to reduce if the MPA threshold is increased (e.g. -13p per week for a 65p MPA, a 0.4% reduction) and increase if the MPA threshold is reduced or removed (e.g. +5p per week, a 0.2% increase if MPA is removed). These patterns are similar across all socioeconomic groups, although with larger changes in spending in more deprived groups (e.g. -23p per week in the most deprived vs -6p in the least deprived quintile under a 65p MPA).

Government revenue from alcohol taxes is estimated to fall if the MPA threshold is increased (e.g. -£109m per year for a 65p MPA), with the majority of this coming from falling revenue from off-trade sales (-£95m vs -£14m from the on-trade under a 65p MPA).

An increase in the MPA threshold is estimated to lead to fewer deaths overall, with 902 fewer deaths from all causes over 20 years under a 65p MPA compared to 200 additional deaths over 20 years if MPA is removed.

The impact on alcohol-specific deaths is larger, with a 65p MPA leading to an estimated 628 fewer alcohol-specific deaths over 20 years, a 4.4% reduction.

By far the largest reduction in deaths under an increase in the MPA threshold is estimated to come from the most deprived quintile of the population (389 fewer deaths in the most deprived vs 71 fewer in the least deprived quintile under a 65p MPA). The converse is true for a reduction in or removal of MPA, with a larger increase in deaths in the most deprived groups.

Modelled impacts of changes to the MPA threshold on hospital admissions follow similar patterns, with higher thresholds leading to larger falls in admissions (e.g. -7,270 over 20 years for a 65p MPA) and reductions in or removal of MPA leading to a rise in admissions (e.g. +1,830 over 20 years if MPA is removed entirely). As with deaths, the largest impacts are in the most deprived groups, with over 4 times as many admissions averted in the most vs the least deprived quintile under a 65p MPA and 4.5 times as many additional admissions in the most vs the least deprived quintiles if MPA is removed.

Patterns of estimated impacts on years of life lost to premature death are similar again – larger reductions for higher MPA thresholds and bigger impacts in more deprived groups.

Increasing the MPA threshold is estimated to save the NHS a substantial amount, with a 65p MPA leading to a saving of £14.9m over 20 years, while removing MPA entirely would cost the NHS £3.8m over the same period.

Comparison of changes to the MPA threshold with changes in alcohol duty rates

We estimate that to achieve the same reduction in alcohol-specific deaths as an increase in the MPA threshold to 65p, alcohol duties would have to rise by 5.8%. Conversely, removing MPA entirely is estimated to have the same impact on alcohol-specific deaths as a 2% cut in duty rates.

As MPA only affects the prices of the cheapest products, whereas duty changes affect all products in the market, larger changes in duty rates are required to achieve the same change in alcohol-specific deaths in the most deprived quintile as a change in the MPA threshold. A 7.4% increase in duties is estimated to achieve the same impact as a 65p MPA in this group, while a 2.8% cut in duty has the same impact as removing MPA entirely.

This more targeted nature of MPA compared to duty means that for the same impact on deaths, an increase in duty rates leads to a substantially larger reduction in consumption in the least deprived quintile and a smaller reduction in the most deprived quintile, compared to increasing the MPA threshold.

Increasing duty rates is estimated to lead to increases in government revenue (e.g. +£35m for a 5.8% duty rise) compared to reductions under MPA (e.g. -£109m for a 65p MPA).

As with impacts on consumption, the targeted nature of MPA means that a duty increase achieving the same reduction in population level deaths as a change in the MPA threshold is estimated to lead to a larger reduction in deaths in the least deprived quintile and a smaller reduction in the most deprived quintile compared to MPA. Both policies are estimated to reduce health inequalities, but MPA does so to a greater extent (e.g. a 65p MPA reduces deaths in the least deprived quintile by 2.4% and the most deprived quintile by 5.5%, while a 5.8% duty increase reduces deaths by 3.6% and 3.9% respectively).

Modelled impacts of alternative approaches to uprating the MPA threshold in future

Adjusting for inflation, we estimate that recent high levels of inflation mean that, by 2026, the real-terms value of a 50p MPA threshold will have fallen to the equivalent of 39p in 2020. Equivalently, the MPA threshold in 2026 would need to increase to 65p per unit to maintain the same real-terms level as 50p per unit in 2020.

If the MPA threshold was increased to 65p in 2026, then maintained at that level without further adjustment, we estimate that alcohol consumption in 2045 would be 2.1% higher and there would be 4,981 additional harmful drinkers in Wales, compared to if the MPA threshold was increased in line with CPIH inflation each year (i.e. maintained at the same level in real terms).

Alternative approaches to regularly uprating the MPA threshold – uprating annually or every 5 years in line with RPI or average earnings, are estimated to reduce alcohol consumption relative to annual CPIH uprating as they would lead to higher MPA thresholds. This is particularly true for indexing the MPA threshold to average earnings, as these typically increase above inflation.

Failing to uprate the MPA threshold in line with inflation is estimated to lead to the largest increases in consumption in the most deprived (+2.6%) compared to the least deprived (+1.2%) quintiles of the population. 

Uprating the MPA threshold is estimated to reduce government revenue compared to no uprating, with total exchequer revenue £174m lower over 20 years.

Maintaining the MPA threshold at the same level in cash terms (i.e. no uprating) is estimated to lead to an additional 501 deaths over 20 years compared to annual CPIH uprating, with the largest impact in the most deprived quintile (187 additional deaths).

Failing to uprate the MPA threshold in line with inflation is also estimated to lead to an additional 16,249 Years of Life Lost to premature death, 3,889 hospital admissions and an additional cost of £7.9m to the NHS over 20 years, compared to annual uprating.

Estimates from an updated version of the Sheffield Tobacco and Alcohol Policy Model for Wales suggest that:

Increasing the MPA threshold from its current 50p/unit level would lead to further reductions in alcohol consumption and harm, with the biggest impacts in the most deprived groups.

Reducing the MPA threshold, or removing MPA entirely, would increase alcohol consumption and harms, increasing health inequalities.

Significant increases in alcohol duty rates would be required to achieve the same reductions in alcohol-specific deaths as an increase in the MPA level and increasing duty does not reduce health inequalities to the same extent.

The implementation of a mechanism to uprate the MPA threshold in line with inflation is important to prevent alcohol consumption and harms increasing as the real-terms value of the MPA is eroded over time.

[1] Moderate drinkers are those drinking within the UK low risk guidelines of 14 units per week. Hazardous drinkers are those exceeding these levels but drinking no more than 35 units per week for women or 50 units per week for men. Harmful drinkers are those exceeding these levels.

Report author: Morris, Gillespie, Kai Le Chen, Wilson, Brennan, Holmes and Angus (2025)

Views expressed in this report are those of the researchers and not necessarily those of the Welsh Government.

For further information please contact:
Social Research and Information Division
Knowledge and Analytical Services
Welsh Government
Cathays Park
Cardiff
CF10 3NQ

Email: Research.HealthAndSocialServices@gov.wales

Social research number: 81/2025
Digital ISBN: 978-1-80633-133-8