Floor area is not one measurement

A price per square metre is a division, and almost all of the disagreement between published rates is in the denominator rather than the numerator. Prices are a matter of record. Floor areas are a matter of convention, and the conventions differ by country, by purpose and sometimes by who is holding the tape measure.

What does a floor-area convention have to decide?

Given a dwelling, which of these counts towards its area? Internal walls. External walls. A sloping ceiling under 1.8 metres. A balcony. A terrace. A cellar. A loft you can stand up in. An integral garage. A shared hallway. A lift shaft passing through the flat.

There is no natural answer. Each convention picks a set because of what it was designed for: a tax base, a sale contract, an energy assessment, a construction cost estimate. A measurement designed to protect a buyer from overstatement will exclude more than one designed to describe a building’s heat loss.

How do the three registers measure floor area?

None of those three was designed to support a price per square metre, and that is the general case rather than an accident: floor areas are collected for administrative reasons and borrowed afterwards by anyone computing a rate.

How many definitions does one country have?

More than one, which is the part that surprises people. France alone has three legal floor areas for a dwelling and they are not interchangeable, so the same flat can be advertised at three different sizes without anybody lying.

• Superficie privative, the loi Carrez measure, defined at article 4-1 of décret n° 67-223: the floors of enclosed and covered rooms after deducting walls, partitions, stairs, ducts and openings, ignoring anything under 1.80 m of height. It applies only to a lot in a co-ownership, and article 4-2 excludes lots under 8 m². This is the contractual figure, and an error over five per cent gives the buyer a claim.
• Surface habitable, at article R156-1 of the construction code: the same structural deductions, plus the exclusion of unconverted attics, cellars, basements, garages, terraces, loggias, balconies, verandas and outbuildings. It is the smaller of the two, and it is the figure a residential tenancy uses.
• Surface utile, which adds back a share of those annexes and is therefore the largest of the three. It is the reference for subsidised housing and several tax reliefs.

None of these is the number in the transfer register, which records a cadastral built area. So the French price per square metre published anywhere, including on this network, divides by a fourth measurement that no advertisement quotes. The French market piece sets all four out with the articles that define them.

England and Singapore are simpler only because each has one dominant convention rather than because anyone chose well: an EPC area exists to model energy use, and a housing authority’s area exists to administer a flat. Neither was designed to be a denominator, and both are used as one everywhere.

What does that mean when you compare two rates?

Within a market, a rate is comparable. Every home in England & Wales is measured by the same convention for the same reason, so a rate in one district can be set against a rate in another. That is the comparison the rate is for, and it is a good one.

Across markets, it is not. A French rate and an English one are the same arithmetic over two different denominators. The difference is not a rounding error and it does not average out, because it is systematic: one convention consistently counts things the other does not.

Within a market, the convention still has limits. An energy assessor measures the dwelling as it stands on the day of the assessment, so a home extended after its certificate is priced at its old size. What an EPC floor area includes and misses is the England & Wales version of that question in full.

How do you tell which convention a figure used?

Before comparing two rates, ask whether the same person measuring the same house would get the same number under both conventions. Inside one country the answer is broadly yes. Across two it is broadly no, and the size of the discrepancy is unknown rather than small.

That is one of the reasons this network publishes each market separately and does not produce a cross-country table: the denominator problem is the second of five obstacles and it is the one that adding a market makes worse rather than better.