Baseline Drift and Schedule Slippage in the UK Government Major Projects Portfolio, 2012–2026

Public reporting on major project performance is usually read as evidence about cost overrun. This study assembles a project-year panel from the annual Government Major Projects Portfolio (GMPP) data published by the United Kingdom government between 2012 and 2026 — 2,532 project-year records across fourteen snapshots, reconciled project-for-project against the official portfolio totals — and shows that the published fields do not support an overrun measure at all. No release, in any year, publishes a forecast whole-life cost; only an approved baseline is published. What can be measured is baseline drift, the year-on-year movement of that approved figure, which is a lower bound on overrun. Across the six snapshot pairs with reliable project linkage, median annualised baseline drift is 0.0 per cent while the mean is +53.1 per cent, and median schedule slippage is zero months against a mean of +4.6 months. The choice between median and mean, computed from the same column, reverses the conclusion. Two further results follow. The share of projects publishing a whole-life cost fell by exactly twenty percentage points, from 90.4 per cent in 2019 to 70.4 per cent in 2026; a balanced-panel decomposition attributes this to composition rather than to progressive redaction, since projects entering the portfolio disclose at 75.9 per cent while those leaving disclosed at 91.2 per cent. Delivery Confidence Assessments are strongly autoregressive but do not detectably predict subsequent drift or slippage. Five structural breaks in the published series are documented, together with the procedure required to survive them.