Design quality is hard to define and harder to measure, so the market has settled on a working proxy: prime, meaning newly built or comprehensively refurbished space. On that definition the evidence is unambiguous, and it points at a value-creation strategy rather than a valuation curiosity.
Start with what is already settled, because it is not the interesting part. A large literature shows that environmental certification carries a measurable premium: roughly 3 per cent on rents and 16 per cent on prices in the original US study of Energy Star and LEED buildings [1]; 19.7 per cent on rents and 14.7 per cent on sales for BREEAM-certified London offices in the 2000s [2]; about 6 per cent on rents across 7,500 leases in certified continental European buildings once location, size and age are controlled for [3]. Where minimum standards bite, the effect shows up as a discount on the worst stock rather than a premium on the best: 18 per cent lower total returns for EPC F and G assets in the UK in 2020, 14 per cent in the Netherlands, with no equivalent effect in European markets that had not yet legislated [3].
Two conclusions follow, and both are now consensus in the industry. Certification is a licence to trade rather than a source of edge, and the effect erodes as certified supply grows: each additional certified building in a neighbourhood reduced the marginal effect of certification by 2 per cent on rents and 5 per cent on transactions [2]. Savills, looking at the stable 38 per cent prime rental premium in Europe, draws the same inference, that "a brown discount is more observable than a green premium" [4].
For an investor, energy performance is therefore an objective of a refurbishment, not a reason for one. The question worth asking is whether the design of the building, the part that cannot be certified, is priced too.
Less than the industry pretends. The literature that isolates architectural quality in offices is small, old and almost entirely American.
The best-known study found that US offices designed by Pritzker Prize or AIA Gold Medal winners let at 5 to 7 per cent above comparable buildings in the same submarket and sold for 17 per cent more, though the authors' own second-stage model confirmed only the rental premium and they cautioned that micro-market conditions might explain it [5]. An earlier study of 102 Class A offices in Boston and Cambridge found that buildings in the top quintile of assessed design quality earned almost 22 per cent higher rents than the bottom quintile, while warning that good design also costs more to produce, so it "may not in fact be more profitable on average, but as with a lottery, may provide a small probability of a high return to the developer" [6].
Where design is decomposed into attributes that can be measured, the findings are more useful. Daylight availability alone carries a 5 to 6 per cent rent premium across 5,145 Manhattan office spaces [7]. In the City of London, a public terrace or rooftop is worth roughly £6.50 per square foot a year, while conferencing facilities, on-site gyms and touch-down space carry no measurable premium at all [8]. Wellbeing certification, which is closer to design than to energy, shows a 4.4 to 7.7 per cent effective rent premium in ten US cities, and that result is explicitly independent of LEED certification, age, renovation and submarket [9].
Knight Frank reached the same place from the opposite direction. Across London retrofits, schemes achieving above-average rental uplift averaged 5.6 amenities against 5.0 for those below: the count barely separates them. What separates them is which amenities. Some 70 per cent of the outperformers had outdoor space against just over 40 per cent of the underperformers, while gyms and car parking showed little effect [10]. An academic hedonic study of 2004 to 2020 leases and an agency analysis of 2020 to 2024 retrofits, using different methods, agree: specification checklists do not price, design judgement does.
There is no European study isolating a design-quality premium in offices, and none linking architectural awards to office rents. That gap is worth stating plainly, because it is the reason the market uses a proxy.
"Prime" is not a marketing term. Knight Frank defines prime space as "new or comprehensively refurbished, rich in amenities and in the most desirable locations" [11]. J.P. Morgan Asset Management defines a prime yield as the yield on "a fully let grade A building in a prime location" [12]. Both definitions bundle design, specification, amenity and location into one observable category. It is imperfect, because location is inside the bundle and location dominates: DWS notes that "average-quality office assets in the best submarkets are outperforming best-in-class office assets in weaker, less desired submarkets" [13], and Savills finds each five minutes closer on foot to a major transport hub is worth about 6.7 per cent of rent [4]. But it is measurable, it is priced daily, and every serious research house reports it.
Rents have separated. Savills puts average European prime rents up 27 per cent since end-2019 against 9 per cent for secondary CBD space, and in the City of London prime is up 49 per cent while secondary has fallen 19 per cent [4]. DWS puts the prime-to-secondary rent spread in European core markets at roughly 35 per cent [13].
Vacancy tells the same story inverted. Headline European vacancy is 9.4 per cent, CBD vacancy 4.9 per cent, prime CBD vacancy around 2 per cent [4]. In central London the overall rate is 9.1 per cent while prime is 2.6 per cent, and 0.3 per cent in the West End Core [11]. In Milan the overall rate is 9.2 per cent, Grade A 3.6 per cent, and 0.6 per cent in the Porta Nuova CBD [14]. There is no shortage of offices. There is an acute shortage of the offices occupiers want.
Capital markets price the difference. Knight Frank's June 2026 yield guide puts prime regional UK offices at 6.50 per cent against 11.00 per cent or above for secondary, and South East towns at 7.25 per cent against 11.50 per cent or above [15]. The comparison that matters is with another sector: in the same guide, prime distribution warehousing sits at 5.25 per cent against 6.00 to 6.25 per cent secondary, a spread of roughly 100 basis points against more than 450 in offices. This is a quality repricing specific to offices, not a general flight from risk.
Green Street, whose Pan-European index deliberately tracks "average institutional quality properties", has been explicit about the split, describing a "bifurcation in favour of 'A' space that began to assert itself post-pandemic" and accelerated through 2023, "eroding further fundamentals for 'B'-quality" [16]. Their conclusion on what to do about it is the same as ours: "developers are likely to find better risk-adjusted returns on offer by pursuing major refurbishment and/or redevelopment works relative to ground-up new construction projects" [16]. Barclays, in March 2026, put the demand side bluntly: demand for the best space is increasing "whereas the worst space is effectively obsolete" [17].
Grade A space accounted for around 92 per cent of central London take-up in the first quarter of 2026 [18] and 70 per cent of Milan take-up in the first half [14]. That is not a preference, it is the market.
The clearest evidence that better buildings let faster is the pre-letting record. Of the 1.84 million square feet completing in central London in the second quarter of 2026, 93.4 per cent had been pre-let before practical completion [19]. Across a record 8.5 million square foot delivery year, around two thirds was already committed [18]. Knight Frank's retrofit study puts numbers on the gradient within the prime tier itself: London refurbishments taken to the top of the quality curve pre-let on average nearly six months before completion, against just over two months for the tier below; for new builds the gap is 14 months against two; and leases on the best buildings run 8.5 years, more than a year longer [10].
Individual schemes make the point concretely. GPE's 2 Aldermanbury Square, 321,650 square feet, was "pre-let entirely off-plan to Clifford Chance" [20]. Derwent London pre-let all of Network W1 to Databricks shortly before completion, at 5 per cent above the December 2025 estimated rental value and 22 per cent above underwriting [21]. British Land's Broadgate Tower, a major refurbishment rather than a new build, was 59 per cent let or under offer more than a year before completion [22].
Supply will not close the gap quickly. London construction starts fell 35 per cent in 2025, new builds more than halved, refurbishments were two thirds of what did start, and Deloitte points to a supply gap from 2027 to 2030 [23]. British Land estimates a 10.4 million square foot shortfall of new or substantially refurbished space in London to 2030 [22].
Two caveats keep the thesis honest.
The first is that location remains the dominant variable, as DWS and Savills both show [4][13]. A design-led refurbishment in a weak submarket is a design-led refurbishment in a weak submarket.
The second is that the prime premium is currently expressed in income, not yet in value. Landsec reported estimated rental value growth of 7.1 per cent for the year to March 2026, its highest in a decade, and still recorded a 1.6 per cent fall in office-led valuations because yields moved against it [24]. Gecina reported values broadly stable, "mirroring market polarization", with a negative yield effect of 1.5 per cent [25]. Prime income is compounding and secondary income is not, but capital markets have not yet paid for the difference.
For a buyer rather than a holder, that is the opportunity rather than the objection.
In a bifurcated market, acquiring obsolete offices in strong locations and redesigning, retrofitting and refurbishing them to prime is a clear value-creation strategy, and it is available now because the pricing gap sits between what secondary assets cost and what prime assets earn.
The mechanism is not the certificate. It is that beautifully redesigned, biophilic, sustainable, amenity-rich buildings in good locations are more attractive to occupiers, and therefore let faster, often before practical completion, at higher rents and to better-quality tenants, than buildings whose owners did not invest in design quality, amenity and sustainability. Every link in that chain is now visible in the data: 92 per cent of take-up going to Grade A, 93 per cent of new completions pre-let, six to fourteen months of pre-letting lead time for the best product, leases a year longer.
Savills' own calculation is the clearest statement of the arithmetic: the payback period for a landlord to take a secondary CBD building to prime through comprehensive refurbishment "has fallen from ten years to five" [4]. Read against secondary yields above 11 per cent, that is the case for our strategy, made by other people's numbers.
This article is provided for information only and does not constitute investment advice or an offer or solicitation. Figures are drawn from the sources cited and were current at the time of writing.