Skip to main content

Quarterly research dataset

The PayslipIQ Anomaly Index

How often are UK payslips wrong? We aggregate anonymised data from every PayslipIQ check and publish quarterly. Free to cite, free to republish, free to challenge.

Press contact

For data requests, methodology questions, or interview pitches: anomaly-index@payslipiq.co.uk

Embargoed advance copies of the next quarterly report are available to UK personal-finance journalists on request.

Q2 2026 — Headline findings

Based on payslips checked through the PayslipIQ tool between April and June 2026. All figures are aggregated and anonymised; we never store individual payslip data.

0.7%

of UK payslips checked had an unresolved emergency tax code (W1/M1/X) more than two pay periods after a job start

7.2%

had a non-standard tax code suffix (K, T, BR, D0, D1, NT, 0T) where the worker reported only one job

2.1%

failed our reconciliation check — gross minus declared deductions did not equal net pay within £1 tolerance

National extrapolation

Applying the percentages above to the 32.7 million UK PAYE workers (HMRC, 2026/27):

AnomalyEstimated workers affectedIndicative annual £ at risk
Unresolved emergency tax code~229,000£137 million*
Non-standard code despite single job~2.35 millionInvestigation warranted
Arithmetic / reconciliation mismatch~687,000Variable, both directions

*Indicative figure. Calculated as 229,000 workers × £600 average overtaxation per emergency-coded year. Actual figure will vary based on individual circumstances; methodology is transparent below.

Methodology

  1. Every payslip submitted to PayslipIQ is processed in-memory by Anthropic Claude Sonnet 4.6 vision; the original image is discarded immediately after the response is returned.
  2. Aggregate flags are written to a counter only — no payslip content, no personal data, no employer data.
  3. We publish only flag types whose national prevalence we are confident in to within ±15%, based on a minimum sample size of 5,000 checks per quarter.
  4. National extrapolation uses HMRC's published figure of 32.7 million PAYE workers in tax year 2026/27.
  5. Average overtaxation estimates are drawn from LITRG and HMRC published figures.
  6. Sample bias caveat: PayslipIQ users are self-selecting (they suspected a problem). Real population prevalence is likely lower, but the size of the discrepancy is what matters for journalistic interest.

Why this matters

UK PAYE is the world's most automated income-tax system. It is also one of the most opaque to the people inside it. Most workers have no easy way to verify their tax code is right or that their employer's payroll has matched HMRC's record.

The Anomaly Index exists to add public-interest pressure on three fronts: HMRC dynamic coding accuracy, employer payroll quality, and worker financial literacy. We will publish quarterly. We will publish the methodology in full. We will respond to peer challenge.

Worried your payslip is one of them?

PayslipIQ checks any UK payslip in 30 seconds. Free, in-memory, never stored.

Check my payslip →

Cite this dataset

PayslipIQ. (2026, Q2). The PayslipIQ Anomaly Index: UK Payslip Errors Q2 2026.
Retrieved from https://payslipiq.co.uk/anomaly-index
Dataset licensed CC BY 4.0.

Disclaimer: The Anomaly Index is a research dataset published for public-interest purposes. It is not a substitute for professional tax advice. Indicative £ figures are estimates based on documented sources and clearly described assumptions. Where you reproduce these figures, please cite the source and the methodology page.

Check Your Payslip Free