Definition
The vouching standard refers to the quality and contemporaneity of records a company must hold to defend a staff-cost or time-apportionment figure in an R&D claim. HMRC does not prescribe a specific timesheet format, but the CIRD manual at CIRD81700 makes clear that apportionment must be based on a reasonable estimate made close to the time the work was performed. Retrospective reconstruction from project names alone is unlikely to satisfy the standard. The test is whether a tribunal, if asked, would find the figures credible in light of the supporting documents.
What records meet the standard
Records that typically satisfy HMRC include: weekly or monthly timesheets signed by the employee and their manager; project management tool exports (Jira, Linear, Azure DevOps, GitHub Issues) showing time logged to qualifying projects; payroll records corroborating the gross salary figures used; and diary or calendar evidence where specific meetings or sprints can be mapped to R&D activity.
Records that typically do not satisfy the standard include: a year-end spreadsheet prepared by the finance team assigning percentages without underlying contemporaneous data; percentages that round uniformly to 50% or 80% across all developers; and testimonial letters from the technical director written after an enquiry has opened.
How HMRC applies it in practice
During a compliance check, an HMRC officer will typically request the underlying time records that support the staff-cost figures submitted in the Additional Information Form. Where records are inadequate, HMRC may seek to reduce the qualifying staff percentage or disallow the cost category. In severe cases, where the apportionment appears to have been invented rather than estimated in good faith, HMRC may raise a penalty.
The Hadee Engineering case at the First-tier Tribunal ([2020] UKFTT 497 (TC)) illustrates what happens when contemporaneous records are absent. The tribunal found the taxpayer's evidence of R&D activity unconvincing partly because the staff time allocation could not be corroborated by records created at the time.
Common mistakes
The most common mistake is preparing the time apportionment only when the accounts are being filed, using memory or project names rather than records from during the year. A second mistake is applying a single percentage to all developers regardless of how their actual work differed across projects. A third mistake is confusing the time apportionment methodology with the vouching standard: having a reasonable methodology is different from having evidence that the methodology was actually applied contemporaneously.
If you are preparing a claim and want to check whether your current record-keeping would withstand scrutiny, the free assessment process covers this point.