The economics of height: rent premium, cost curve, cap rates

TL;DR

• Rent rises roughly 0.6% for each floor above the second, and faster above the fortieth. Cost rises far quicker: for tall residential buildings the estimated height elasticity of construction cost is about eight times the height elasticity of rent.
• The tower also keeps less of itself. Office and mixed-use supertalls average below 72% net-to-gross, and each additional metre of height is associated with a 0.078% fall in space efficiency.
• Capital is now the binding constraint. European all-in debt cost reached 4.7% in Q1 2026 against a 4.9% average prime office yield, and Knight Frank reports that London's largest assets trade at a yield discount, not a premium.

Height earns about 0.6% more rent per floor. Building it costs several times that. Every tower in Europe is a bet that the gap can be closed by location, scarcity and a long lease, and in 2026 that bet is harder to win than at any point since the last cycle. The arithmetic below sits underneath why height still matters, and underneath every scheme quietly shelved in the past 18 months.

The rent gradient is real, and it is shallow

Tenants do pay for altitude. The largest study of within-building rents, by Crocker Liu, Stuart Rosenthal and William Strange in the Journal of Urban Economics, found the gradient is not a straight line. Rent drops roughly 50% moving from the ground floor to the second, because retail access is worth more than any view. Above the second floor rent then climbs about 0.6% per floor, and the curve steepens high up, above floor forty. In their published estimates, moving up one floor lifts rent by about as much as adding 3,500 workers to the building's zipcode.

Amsterdam evidence explains what the money buys. Nase, van Assendelft and Remøy modelled 627 office lease transactions across 33 tall Amsterdam buildings between 2000 and 2016, and decomposed the vertical premium: 27% relates to view, 3% to industry-level differences, and the remaining 70% to firm-level signalling and other factors. Only a quarter of the premium is the window. Most of it is the address, which is why the tenants profiled in who leases the top floors behave the way they do.

Stack the gradient and a fortieth floor earns meaningfully more than a third floor in the same building. It does not earn double.

The cost of height outruns the rent it buys

Gabriel Ahlfeldt and Jason Barr put numbers on both sides of the trade. They estimate the height elasticity of per-unit construction cost at 0.1 for buildings up to nine floors. Beyond that the cost of height increases sharply: 0.25 for tall commercial buildings and 0.56 for tall residential, roughly twice as steep, because flats need windows in more rooms, smaller floor plates, more plumbing and more complicated facades.

Against that they estimate the height elasticity of per-unit floor space rent at about 0.07 for residential units in New York and Chicago. Compare like with like and the residential cost elasticity is around eight times the residential rent elasticity. That single ratio is the reason the European residential tower question keeps resolving against the tower.

Jason Barr's Manhattan sample says the same thing in dollars. Across 28 projects built since 2001, inflation-adjusted to 2016, construction cost ran near $400 per sq ft for a 20-storey building and near $800 per sq ft at 80 storeys.

Then the floor plate shrinks

The cost side has a second mechanism, and it is the one most pro formas underweight. Cores, lift banks, refuge floors and plant levels consume area nobody pays rent on.

Now more than ever, it is important to recognise that shape is as important as height when it comes to delivering a project cost efficiently.

Steve Watts, Turner & Townsend

Hüseyin Emre Ilgın's study of 166 tall and supertall towers, published in the Journal of Sustainability Research in September 2025, quantifies the loss. Hotels average 81.2% net-to-gross and residential towers about 77%, while office and mixed-use supertalls fall below 72%. Form matters as much as use: prismatic towers average 75.1%, free-form 74.0% and tapered 72.0%. Across the dataset, each additional metre of height is associated with a 0.078% decline in space efficiency.

Read those two findings together. Rent is earned per square metre of lettable area, and height removes the square metres it is supposed to be paid for.

London's cost curve bent 40% in five years

The cost curve moved sharply after 2020. Turner & Townsend's Tall Buildings Construction Guide 2026, published in March 2026, put the average cost of building a tower in London at £3,800 to £4,300 per sqm in 2020 and £5,200 to £6,500 per sqm in 2025, a rise of up to 40%. New York and Seoul rose about 30% over the same period, Tokyo 35%.

The absolute spread across markets is wider still. Office shell and core in 2025/26 runs $6,900 to $8,700 per sqm GIA in London and $7,000 to $9,700 in New York, against $1,500 to $1,800 in Dubai and $705 to $805 in Mumbai. London residential shell and core runs $4,900 to $5,700. That gap is a large part of why the pipeline covered in Dubai as a proving ground behaves nothing like London's.

Geometry is the controllable variable. Turner & Townsend price facade at roughly $47 per sq ft at a wall-to-floor ratio of 0.30, rising to about $95 per sq ft at 0.60, and put the spread between London's most ambitious and most cost-efficient schemes at 25%. Steve Watts, the firm's tall buildings lead, frames the consequence directly: shape is as important as height in delivering a project cost efficiently. On viability he is blunter, calling it the most pressing issue facing the sector.

Who gains, who pays

Winners:

• Owners of standing prime towers with income, who benefit from a development pipeline that has stopped. London completed 16 high-rise buildings in 2019 and is projected to deliver one in 2026, the slowest year since 2015.
• Occupiers already inside a tower on a pre-2021 lease, before City rents doubled.
• Buyers writing tickets below £100m, the size band Knight Frank identifies as clearing best.

Losers:

• Developers underwriting new starts, who need more than £105 per sq ft, in James Abrahams of Allsop's phrase, "just to put a spade in the ground".
• Vendors of very large single assets, where the equity pool thins out fastest.
• Residential tower promoters, carrying the 0.56 cost elasticity against a rent elasticity near 0.07.

Towers trade at a discount, not a premium

The trophy premium investors assume is not visible in the price. In December 2025 Nuveen sold 70 St Mary Axe, the 21-storey City of London tower known as the Can of Ham, to Hayfin and Capreon. The parties did not disclose the price. Property Week reported £331m, or £1,049 per sq ft, at a net yield of 5.81%. The building opened in 2019, is fully let to 13 tenants, and carries debt from Santander and CaixaBank.

That 5.81% sits well above the roughly 3.75% prime West End yield. Shabab Qadar of Knight Frank explains the pattern: the best London transactions are happening below the £100m mark, often at implied yields inside the 5.25% average, so larger assets now trade at a yield discount. Size is a penalty, not a premium.

Warsaw prices the same way. Ghelamco sold the 202-metre, 46-floor Warsaw UNIT to Eastnine in November 2024 at an underlying property value of €280m against €18.0m of annual rent, an income yield of 6.4%, with a WAULT of 5.1 years. By March 2026 Cushman & Wakefield put prime Warsaw office yields at 6.00%. The city's tallest office asset cleared at roughly the market's prime rate, not inside it, which is the yield reality underneath how a mid-size capital built a skyline.

The best transactions in London at the moment are happening below the £100m mark. When you look at the implied yield, they're often below the [average] of 5.25%.

Shabab Qadar, Knight Frank

Cushman's Warsaw numbers also show where the real premium lives. Prime central rents stood at €24 to €29 per sqm per month in March 2026 against €15 to €19 in non-central locations, with central vacancy at 6.5% versus 12.2% outside. Location moves rent by more than half. Height moves it by 0.6% a floor.

Debt now costs what the building yields

The financing test is arithmetic. Savills reports that the average all-in cost of debt across selected European markets rose 60 basis points in Q1 2026 to 4.7%, which the firm describes as now in line with the average prime office yield of 4.9%. Loan-to-value sits at 55% for prime continental offices and 60% for London City offices.

When debt costs what the asset yields, leverage stops adding return and a large ticket becomes an equity problem. The German market shows the effect at scale: office investment volumes reached only €700m in Q1 2026, down 82% on the ten-year average.

The occupier side, by contrast, keeps improving. Allsop's Abrahams reports £150 per sq ft achieved in a City tower, against the mid-£70s when 22 Bishopsgate completed in December 2020. Knight Frank recorded prime City core rent growth of 7.9% in 2025 and puts London office investment at £9.3bn for the year, forecasting £12bn in 2026.

What to watch, and when

The height premium is not broken. It is narrower than the cost curve, and for now the debt curve is sitting on top of both.

Three dates decide the next move. Savills expects the European Central Bank to raise rates 50 basis points to 2.50% in summer 2026 before cutting back to 2.00% in 2027, which delays prime yield compression across most jurisdictions until 2027. Savills also expects higher steel prices and tariffs to push the development pipeline back until well into 2027 as developer margins are squeezed. Watch the Q4 2026 and Q1 2027 prints for the first quarter in which prime yields move out ahead of debt costs again, and watch whether any European tower start is underwritten before that crossover.

Until it happens, capital has an easier answer than building. Over €3bn of continental European office stock was bought for redevelopment in 2025, the strongest year since 2021, which is why the sharper European question has moved from what to build next to what to do with the towers that already exist, and why London's tall pipeline is thinner than its planning consents suggest.

More from the Skyscraper Day Dive.