Between the Brexit referendum (June 2016) and the onset of COVID (early 2020), London property markets produced some of the dullest headline numbers in living memory. The London-wide median transaction price per square metre rose from £5,761/sqm to £5,829/sqm over three and a half years — a nominal gain of just +1.2%. For a city that had tripled in price over the previous two decades, this felt like stagnation.
It was not stagnation. Our hedonic factor model — which strips out shifts in the mix of properties transacting and isolates quality-adjusted price changes — tells a different story. And beneath the flat headline, one factor was quietly undergoing one of its largest moves in the entire dataset.
The headline median disguises compositional effects: the mix of what sold changed. Controlling for property type, size, location, energy rating and construction period, the baseline market factor rose +13.7% during this period. That is not spectacular, but it is not stagnation either — it is broadly in line with general inflation over the same period. The market was treading water in real terms, not falling.
The gap between the +1.2% nominal print and the +13.7% quality-adjusted number is explained by a shift in transaction mix: more small flats, fewer large houses, more outer-London activity — all of which depress the simple median even when underlying values are rising.
The most significant factor move of this period — and a foreshadowing of what came next — was the erosion of the flat premium. In mid-2016 flats commanded a +10.5% premium over equivalent houses (same postcode, same floor area, same energy rating). By October 2019 that had contracted to +5.9%: a fall of 4.9%.
Given that London's market is disproportionately composed of flats, particularly in inner zones, this relative underperformance was a direct drag on the widely-watched average price indices. Owners of houses, especially in outer London, had a materially better experience during these years than owners of flats — even if they never saw it in the headlines.
The new-build factor also deteriorated modestly. New builds moved from roughly par to a discount of -1.1% relative to equivalent second-hand stock during this period. Help to Buy was sustaining volumes but not protecting resale values relative to the broader market.
| Factor | Period return | Direction |
|---|---|---|
| Baseline Market | +13.7% | Quality-adjusted prices rose despite flat nominal |
| Flat Premium | -4.9% | Flats lost ground vs equivalent houses |
| Floor Area | -3.6% | Per-sqm value of extra space began fading |
| Energy Score | -1.1% | Green premium not yet pricing in |
| New Build | -1.1% | Slightly wider discount vs second-hand |
| Freehold | +0.8% | Marginal tailwind for houses |
Factor returns, May 2016 – October 2019
If you owned a house rather than a flat, you experienced roughly +13.7% quality-adjusted appreciation plus a +0.8% freehold tailwind. If you owned a new-build flat, you had the flat discount widen (-4.9%) and the new-build discount widen (-1.1%) simultaneously — a double headwind that offset much of the baseline gain.
The seeds of the post-COVID flat discount were planted here. The flat premium did not collapse overnight in 2022: it had been eroding for at least three years before the pandemic began.
Enter any London postcode to see its quality-adjusted price history, factor contributions, and how it compares to the London average.
Analyse a Postcode →The flat headline of 2016–2019 gave way to something very different: a rapid +14% nominal surge concentrated entirely in outer London, as buyers left Zone 1 in search of space. Read Part 2 →
Factor returns are derived from a rolling 3-month cross-sectional OLS regression of log(price/sqm) on eight property characteristics, run on every Land Registry transaction matched to an EPC certificate. The intercept (Baseline Market) captures quality-adjusted market-wide appreciation; other factors show how the market price of each characteristic evolved. All returns here are computed within this period only. See the factors page for full definitions.