Government Contribution by Ethnicity
Which ethnic groups are net contributors to — or net recipients from — the UK tax and benefit system. Data from the ONS "Effects of Taxes and Benefits on Household Income" report, Table 23.
Five-Year Averages (2019/20 – 2023/24)
Mean values averaged across all five available years.
Net Contribution to Government
Original income minus final income — positive values mean the group pays more in taxes than it receives in benefits. This is calculated by subtracting each group's final income (after all taxes deducted and benefits added) from their original income (before any government intervention). A higher bar means the group is a larger net contributor to the public finances.
Original Income by Ethnicity
Before any taxes or benefits — income from employment, pensions, investments.
Final Income by Ethnicity
After all taxes deducted and benefits (cash + in-kind) added.
A note on the "Mixed" group
The Mixed ethnicity category shows by far the most volatility in this dataset. Its mean original income swings from £41,689 in 2023/24 to £54,019 in 2022/23 — a range of over £12,000 that dwarfs the variation seen in any other group. The same instability appears across all income stages.
This is almost certainly a sample size problem. The ONS Living Costs and Food Survey, which underpins this data, interviews roughly 5,000 households per year. The Mixed ethnic group makes up around 1–2% of the UK population, meaning the estimates for this group may rest on as few as 50–100 households. Small samples are highly sensitive to outliers — a handful of very high or very low earners in any given year can dramatically shift the average.
By contrast, the White group (representing ~80% of the population) draws on thousands of observations and produces much smoother trends. The Asian and Black groups, while smaller, are still several times larger than the Mixed sample and show less year-to-year noise.
The Mixed group data is presented here for completeness, but the year-on-year movements should be interpreted with caution. The five-year average gives a more reliable picture.
How this data is collected
This data comes from the ONS Living Costs and Food Survey (LCF), a continuous household survey that has run in various forms since 1957. Around 5,000 UK households are sampled each year. Each participating household keeps a detailed spending diary for two weeks and provides information on income, taxes, and benefits received.
The ONS then models the full tax and benefit journey for each household: starting with original income (earnings, pensions, investments — what you'd have with no government at all), then adding cash benefits to get gross income, subtracting direct taxes to get disposable income, subtracting indirect taxes (VAT, fuel duty, etc.) to get post-tax income, and finally adding benefits in kind (the estimated value of state education, NHS care, housing subsidies) to arrive at final income.
Incomes are equivalised using the OECD modified scale, which adjusts for household size and composition. A single person living alone and a couple with three children on the same cash income are in very different positions — equivalisation puts them on a comparable footing.
Ethnicity is self-reported by the household reference person. The broad groupings (White, Asian, Black, Mixed, Other) aggregate many distinct communities — "Asian" includes Indian, Pakistani, Bangladeshi, and Chinese households, which have very different income profiles from each other.
What this data tells us — and what it doesn't
It shows the net fiscal position — whether a group, on average, pays more into the system than it takes out, or vice versa. The White group is a consistent net contributor, driven largely by higher original incomes (particularly from employment in higher-paying occupations and from private pensions). The Black group tends to be a net recipient, reflecting lower original incomes that are then substantially topped up by the benefit system.
It shows redistribution working. The gap between ethnic groups narrows significantly between original and final income in every year. The tax and benefit system compresses income differences — that is its design. Groups with lower original incomes receive proportionally more in benefits and benefits in kind (NHS, education), while groups with higher original incomes pay proportionally more in taxes.
It does not tell us why original incomes differ. The gap in original income reflects a tangle of factors: occupational distribution, age profile (younger populations have lower earnings), regional concentration (London inflates some groups' housing costs and earnings simultaneously), household structure (more children = lower equivalised income), qualification levels, language barriers, discrimination, and immigration recency. This data cannot separate those causes.
It does not capture wealth. Income and wealth are different. A retired homeowner with a paid-off house and modest pension income looks "poor" in income terms but may be asset-rich. The LCF captures income flows, not balance sheets.
Benefits in kind are estimated, not observed. The NHS and state education components of final income are allocated based on usage assumptions (e.g. number of children, GP visits), not actual costs. These imputed values are substantial — often £8,000–12,000 per household — and their allocation methodology can significantly affect which groups appear to be net recipients.
The "Other" category is a residual. It includes Arab, any other ethnic group, and those who didn't fit elsewhere. It's a statistically small and very heterogeneous group, making averages particularly unreliable.
How equivalisation distorts the picture
There are two important ways the methodology can mislead when comparing ethnic groups with different demographics.
Larger families appear to receive less than they do. Groups with larger average household sizes — notably Asian households — receive substantial benefits in kind: each child adds roughly £5,000–7,000 of imputed state education value, and more household members draw NHS allocations. However, equivalisation then divides this total across the larger household using a diminishing scale. The benefit spending is real and is counted, but the per-person equivalised figure is dampened. In cash terms, the state may be spending significantly more on a larger household than the equivalised numbers suggest.
Older populations look like bigger net recipients than they may be. The ONS allocates NHS benefit in kind using age and sex-specific cost profiles — older people are assigned far higher values due to greater average healthcare consumption. The White population has a significantly older age profile than most minority groups, meaning White households receive large NHS imputations that inflate their final income and reduce their apparent net contribution. These imputations are based on average costs by age band, not actual individual usage — a healthy 75-year-old receives the same allocation as one with complex care needs. Without this heavy NHS imputation to older White households, the White group's net contribution figure would appear even larger.
Neither of these effects makes the data wrong — equivalisation and benefits-in-kind imputation are the standard methodology. But they mean the headline figures should not be read as a simple "who costs the government more" comparison without understanding what the adjustments are doing beneath the surface.
The missing data
The ethnicity data in this report is equivalised — household income and benefits are divided by a factor that adjusts for household size, then reported per person. This means the figures understate the total government spending received by larger households. If it is true that Asian households are systematically larger than White households (which census data supports), then the real fiscal transfer to Asian households may be substantially higher than these equivalised figures suggest.
The ONS does not publish unequivalised income data broken down by ethnicity. Releasing it would allow a more complete assessment of net fiscal contribution by ethnic group.
About this data
All figures are equivalised household income per year (£) from the ONS "Effects of Taxes and Benefits on Household Income" dataset, Table 23. Equivalisation adjusts for household size and composition.
Ethnicity breakdowns in Table 23 are only available from FYE 2020 (2019/20) onwards. Earlier editions of this dataset do not include income by ethnicity.
Source: ONS — Effects of taxes and benefits on household income ↗