Full Breakdown
Flood Exposure and English House Prices: Findings from a Large-Scale Repeat-Sales Study
6/22/2026, 8:07:26 PM
Study Overview
A research team applied a repeat-sales hedonic price model, extended with an age-period-cohort (APC) framework, to quantify how flooding influences residential property values in England. By comparing sale prices of the same dwelling before and after a flood event, the analysis isolates the price effect of flood exposure from broader market movements.
Data Assembly and Cleaning
The analysis began with the England-and-Wales price-paid dataset. Transactions outside England were removed using English postcodes (January 2023). A logarithmic transformation identified extreme outliers; prices below £5,557.30 or above £3,718,654.30 were excluded, eliminating 19,438 records and leaving 27,758,154 transactions. Linking records by postcode and addressable object names produced 12,009,213 repeat-sale pairings. Spatial attributes were derived from Ordnance Survey Code-Point with Polygons and Local Authority District (LAD) boundaries updated to the 2023 ONS release.
Flood Event Data
Flood exposure was defined using the Environment Agency’s Recorded Flood Outlines dataset, which records fluvial, coastal and sewer flooding. The study retained events from 1 January 1995 onward, producing 78,025 polygon events covering 2,723 km²—89 % fluvial (2,414 km²), 6 % coastal (170 km²) and 5 % sewer (139 km²). The Environment Agency Flood Zone 3 dataset and the Flood Map for Planning (Rivers and Sea) supplied high-risk flood-zone polygons, enabling calculation of the proportion of each postcode district lying within Flood Zone 3 (values from 0 to 1).
Methodological Framework
The repeat-sales hedonic model controls for time-invariant property characteristics, attributing residual price changes to flood exposure and flood history. The APC extension separates three temporal dimensions: age effects (post-flood recovery intervals), period effects (1995–2008 baseline, 2009–2015, post-2016) reflecting macro-economic shocks and policy shifts such as the 2008–09 Global Financial Crisis, the expiry of the Statement of Principles and the introduction of Flood Re, and cohort effects (first recorded flood date in each postcode: pre-2009, 2009–2015, post-2016). Survival analysis, using the Kaplan-Meier estimator from Python lifelines package, estimates the probability that homeowners remain in flood-exposed properties over time.
Findings and Implications
Model coefficients from the hedonic regression quantify the influence of flood-zone coverage and flood history on house prices, while the APC framework separates age-related price dynamics from period-specific shocks (e.g., the 2008–09 Global Financial Crisis, the expiry of the Statement of Principles, the introduction of Flood Re) and cohort-related differences. The survival analysis using the Kaplan-Meier estimator provides estimates of the probability that owners remain in flood-exposed properties over time, offering insight into potential effects on mortgage considerations.
Official Data Sources & Statements
The Environment Agency supplies the Recorded Flood Outlines and Flood Zone 3 datasets, described as the best estimate of historical flood extents in the public domain. The Office for National Statistics provides the 2023 LAD boundaries, and Ordnance Survey delivers postcode-district geometry via the Code-Point with Polygons product. These official sources underpin the spatial and temporal alignment of property transactions with flood exposure.
