The London "flat premium" — how much more a flat costs vs…
0 reactions
The London "flat premium" — how much more a flat costs vs an identical-size house — has collapsed from +10% to +1% today. 30 years of HM Land Registry data.
**Tools:** Python, pandas, statsmodels OLS, matplotlib.
**Data:** HM Land Registry Price Paid Data (\~5M London transactions since 1995) merged by postcode with MHCLG EPC energy certificates.
**Method:** rolling 3-month cross-sectional OLS of log(price/sqm) on hedonic property characteristics (floor area, rooms, EPC band, construction era, flat-vs-house, freehold/leasehold), with postcode-area dummies as controls. The "flat premium" is the coefficient on the flat dummy, how much more per sqm a flat costs vs an otherwise-identical house in the same postcode area.
**What it means:** in May 2023 a London flat was priced \~10% above an equivalent house per sqm. Today that gap is basically zero. This is the post-rate-rise correction expressing itself compositionally, not as a nominal crash.
Full methodology + interactive charts at propertyanalytics.london http://propertyanalytics.london/ .
#education
source
**Tools:** Python, pandas, statsmodels OLS, matplotlib.
**Data:** HM Land Registry Price Paid Data (\~5M London transactions since 1995) merged by postcode with MHCLG EPC energy certificates.
**Method:** rolling 3-month cross-sectional OLS of log(price/sqm) on hedonic property characteristics (floor area, rooms, EPC band, construction era, flat-vs-house, freehold/leasehold), with postcode-area dummies as controls. The "flat premium" is the coefficient on the flat dummy, how much more per sqm a flat costs vs an otherwise-identical house in the same postcode area.
**What it means:** in May 2023 a London flat was priced \~10% above an equivalent house per sqm. Today that gap is basically zero. This is the post-rate-rise correction expressing itself compositionally, not as a nominal crash.
Full methodology + interactive charts at propertyanalytics.london http://propertyanalytics.london/ .
#education
source
| 1,000 mlx | |
| 2 | |
| 0 |