ALIS-Light public report · Office for Budget Responsibility

UK welfare fraud and error: the pandemic spike and subsequent decline

The OBR reports that welfare fraud and error rose sharply during the Covid years and then fell back close to its pre-pandemic rate. The increase was concentrated in Universal Credit, especially among claims that entered during the first year of the pandemic.

Local OBR PDF checked page by page

3.1% -> 4.3% -> 3.2% Overall welfare fraud and error rate

9.4% -> 14.7% -> 8.5% Universal Credit fraud and error rate

25.8% -> 8.9% Covid-cohort UC rate, 2020-21 to 2025-26

Sampling based OBR says the resulting conclusions remain uncertain

Source

Institution: Office for Budget Responsibility (OBR)
Report: Welfare trends report
Publication date: June 2026
Command paper: CP 1634
Evidence used: Local 45-page accessible PDF retained in the report workbench
Trend period discussed here: 2019-20 to 2025-26

Plain-English summary

Across welfare, fraud and error rose from 3.1% of spending in 2019-20 to 4.3% during both 2020-21 and 2021-22. It then fell to 3.2%, close to the pre-pandemic rate.

The movement was concentrated in Universal Credit (UC). The UC rate rose from 9.4% of UC spending in 2019-20 to 14.7% in 2021-22, then fell to 8.5% in 2025-26.

OBR identifies the group joining UC during the first pandemic year as the main driver. That Covid cohort had a fraud and error rate of 25.8% in 2020-21, falling to 8.9% by 2025-26.

The rates at a glance

Measure	2019-20	Pandemic-period high	2025-26
Overall welfare	3.1%	4.3% in 2020-21 and 2021-22	3.2%
Universal Credit	9.4%	14.7% in 2021-22	8.5%
Covid cohort within UC	Not yet formed	25.8% in 2020-21	8.9%

Each percentage is a share of spending for the named scope. Overall welfare, all UC, and the Covid cohort therefore have different denominators.

How OBR interprets the pandemic rise

OBR says the initially high Covid-cohort rate is likely to be primarily explained by operational easements introduced at the start of the pandemic. These aimed to speed access at a time of acute need.

The easements included relaxed departmental checks for claims, suspension of the minimum income floor, and suspension of gainfully-self-employed checks. OBR also points to the sharp increase in self-employed UC cases, which had higher fraud and error rates than the rest of the caseload.

This is OBR's evidence-based judgement, not proof that any one factor caused the whole increase.

Trend interpretation

The data indicate a large but temporary pandemic shock rather than a continuing rise across all welfare benefits. OBR reports that pre-existing UC cohorts were relatively stable while the Covid cohort accounted for most of the spike and much of the subsequent decline.

The later fall is consistent with the withdrawal of pandemic easements, the maturing composition of the Covid cohort, and additional departmental action. OBR still treats the relative contribution of these factors as uncertain.

Change to future forecast methodology

OBR previously assumed that the underlying propensity for fraud and error in UC would rise in every forecast year. Based on the concentration of the increase in the Covid cohort and the return of rates close to pre-pandemic levels, OBR says it will remove that rising-propensity assumption from future forecasts.

This is a planned methodological change for the next forecast. It should not be read as a claim that future fraud and error will necessarily fall.

Limitations and uncertainty

Fraud and error rates are estimated through sampling exercises rather than a complete review of every claim.
OBR says evidence on the underlying drivers is limited, so its causal conclusions remain uncertain.
The percentages measure overpayments as a share of spending; they are not the percentage of claimants committing fraud.
Overall welfare and UC rates cover different spending bases and should not be compared as if they had the same denominator.
The Covid cohort is specifically UC cases joining between 14 March 2020 and 27 February 2021.
Forecast assumptions are judgements and may change as new evidence becomes available.

Verification notes

The source PDF identity, page count, size, and SHA-256 hash are recorded in the workbench.
The required figures were checked on PDF page 8 (printed page 3), PDF page 9 (printed page 4), and PDF page 17 (printed page 12).
The forecast-method statement was checked across PDF pages 10-11 (printed pages 5-6).
All cited pages were rendered and inspected visually in addition to text extraction.

Citation and source

Suggested citation: Office for Budget Responsibility, Welfare trends report, June 2026, CP 1634.

Evidence file: E03599306_OBR_WTR_2026_Accessible_v3.pdf, retained in the GoDataBank report workbench.

Primary report pages: Executive summary pages 3-6 and recent-trends page 12.