The Roman Commuter Belt? Roman Villas and Household Income in England and Wales

Roman villas and today's wealthiest neighbourhoods both lean towards lowland south-east England. Do villa locations still mark the money? I build an open list of 617 Roman villas in England and Wales by joining Wikipedia's lists of villas, Wikidata, OpenStreetMap and the Pleiades gazetteer, and set it against ONS estimates of mean household income for the 7,264 middle layer super output areas (MSOAs) of England and Wales in the financial year ending 2023. MSOAs with a villa within 2 km of their population-weighted centre have a median household income 10.5% higher than the rest, and a regression premium of 8.6%. Comparing within regions cuts the premium to 2.2%; adding distance to London takes it to 0.5%; with controls for Roman roads and towns, elevation, relief, the coast and urbanness it is 0.4% (95% interval −2.0% to +2.8%), and within local authorities 0.3%. A placebo test that moves every villa 5 to 15 km in a random direction produces raw premiums as large as the real one in 94% of 200 runs. At the coarser scale of local authorities, villa density and income are modestly correlated (Spearman 0.27, 0.21 after controls). The villa belt and the commuter belt overlap because both sit in the same broad geography, not because villa sites mark rich places. The villa list is a partial, south-leaning record, and the analysis is correlational. Interactive version: https://tristan-hotham.github.io/roman-villas/. Data and code: https://github.com/Tristan-Hotham/roman-villas. Blog post: https://blog.tristanhotham.com/2026/10/27/roman-villas/.