Lost Money: Roman Coin Loss across England by District and Reece Period, from the Portable Antiquities Scheme

Roman coins are closely dated and common among finds reported by the public in England. I use facet counts from the public search of the Portable Antiquities Scheme (PAS) to build a complete district-level table for England: 275,206 Roman coin records, 188,831 of them assigned to one of Reece's 21 issue periods, set against 1,062,032 records of all periods that measure how intensively each of 296 districts is detected and recorded. The English profile reproduces the classic British site-find curve: loss per year is highest in AD 330 to 348 (4.8 times a flat rate), then AD 364 to 378 and AD 260 to 275, and 82% of dated coins are later than AD 260. I then test, with a plan written before any district shares were computed, whether the share of late coins (AD 364 to 402) rises smoothly towards the south and east and the share of early coins (before AD 222) towards the north and west. Neither gradient is supported: per 100 km east the odds ratio for the late share is 0.94 (95% CI 0.83 to 1.06), and per 100 km north 0.98 (0.89 to 1.08), and all but one of 20 planned checks agree. In exploratory analysis the variation takes the form of a step: in the North East and North West, 4.3% of dated coins are late against 15.9% elsewhere, and 58% are early against 15%. The contrast looks like a frontier zone set against the rest of England, not a continuous tilt.

Interactive version: https://maps.tristanhotham.com/roman-coins/. Data and code: https://github.com/Tristan-Hotham/roman-coins. Blog post: https://blog.tristanhotham.com/2026/10/31/roman-coins/.

Sources: Portable Antiquities Scheme database (finds.org.uk), British Museum and Amgueddfa Cymru, public search facet counts retrieved 2026 (aggregate counts only); Reece period dates from Nomisma.org; ONS Local Authority Districts (December 2023), regions and counties (Open Government Licence v3.0). Text, figures and data CC BY 4.0; code MIT.