The Conversion Trap

I. THE BILLION-DOLLAR PREMISE

There is approximately one billion square feet of underutilized office space in the United States. Developers see distressed assets at deep discounts, cities are dangling tax incentives, and the housing shortage is producing political tailwinds strong enough to streamline approvals. The logic of office-to-residential conversion has never felt more airtight.

And yet, when Goldman Sachs ran the numbers, they found that converting a nonviable office at current market pricing produces a loss of $164 per square foot. Their conclusion: office prices would need to fall by half for the conversion math to work without subsidy.

This is not a financing quirk. It is a structural problem rooted in what the product actually is, what it costs to build, and what it earns. Most of the developers rushing into the space have not paused to distinguish between those variables.

The herd is following each other into a strategy where the cost structure is inverted — paying a premium to produce a product that rents at a discount.

The industry benchmark from Morgan Stanley is unambiguous: acquiring a completed multifamily property costs roughly $600 per square foot. Ground-up development runs about $588. The average cost of acquiring and converting an existing office building lands at $685 per square foot. Conversion, in aggregate, costs more than building new. And it produces a product — standard multifamily rental units — that earns $22 per square foot annually, against the $37 per square foot the building used to generate as office space.

Brookings ran 18 buildings through rigorous financial modeling across six cities. Sixteen of the eighteen showed a negative NPV after conversion costs, even though switching to residential use would increase revenue over dead office. The conversion cost erases the gain.

This is the baseline condition. It is not a fringe outcome. It is the central tendency of the market.

II. WHAT THE BUILDING IS ACTUALLY TELLING YOU

The physical realities of office-to-residential conversion are where the financial picture becomes granular — and where the range of outcomes gets very wide, very fast.

CBRE estimates conversion hard costs at $100 to $500-plus per square foot depending on the building. That range is not ambiguity; it reflects three genuinely different project types.

Pre-war buildings with narrow floor plates and operable windows convert for $150–$250 per square foot. They are the easy ones, and there are not many of them left.

Standard 1970s–90s office buildings run $300–$375 per square foot in hard costs — comparable to new construction in secondary markets, without the flexibility of ground-up design.

Deep-plate, curtain-wall, post-tension buildings cost $500–$600 per square foot or more. These are the ones that look cheap at acquisition and turn out to be the most expensive projects in the portfolio.

The kill factor in the third category is the floor plate. Residential building codes require operable bedroom windows. Most post-war office buildings have floor plates of 100 feet or deeper — far exceeding the 60-foot maximum at which you can reliably put every unit on a perimeter wall. When the floor plate requires carving a central light well, developers lose 30–40% of rentable area. At that loss factor, the project stops working on paper.

There is also the plumbing problem. Unlike office floors where two bathrooms serve an entire tenant, residential conversion requires distributed plumbing throughout. Waste lines slope at a quarter-inch per foot. A 40-foot run drops 10 inches. Stack HVAC ducts and fire suppression, and a 12-foot ceiling — luxurious by office standards — can collapse to 7 feet 9 inches before you finish the ceiling plane. Below 7 feet is non-compliant in most codes. The ceiling is not a design choice at that point; it is a legal constraint.

One striking post-tension cable during core drilling costs $25,000 or more to repair. In buildings where no original structural drawings exist, developers are essentially performing surgery on a patient whose anatomy is unknown.

Kofi Meroe, director at Foulger Pratt and the developer behind Accolade — the 243-unit conversion at 1425 New York Avenue NW near the White House, completed in 2025 — put it plainly: "There is no separate building code for conversions. We have to meet the same building code and performance standards as if I'm building from scratch." The misreading of accessibility standards, elevator placement requirements, or zoning constraints is not a line-item risk. Meroe is explicit: a single missed compliance item can make an already-marginal project unviable.

"Most of these conversion deals are already tight on the yield, so something like missing the right accessibility standards — it can make a project become unviable very quickly." — Kofi Meroe, Foulger Pratt

None of this is fatal to a sophisticated operator. Experienced conversion developers know how to navigate each problem. But the costs of navigation are real, and most pro formas do not model them honestly.

III. THE NOI THAT ISN'T THERE

Even when a conversion is executed well, the revenue picture is stubborn. The multifamily market is structurally oversupplied in the locations where most office conversions are occurring. National vacancy sits at 7.1% — the highest in recent recorded history. Median asking rents dropped 2.2% year over year in late 2024. Lease-up timelines have stretched, concessions are rising, and the Sun Belt markets where conversion activity is most concentrated are among the hardest hit.

The NOI differential between office and residential that is supposed to justify conversion barely exists. CBRE found that in 2022, the average multifamily building at 96.5% occupancy recorded NOI of roughly $16 per square foot. The average office building recorded $15.50. You are taking all the conversion risk, all the construction cost, all the disruption — for a fifty-cent improvement in NOI per square foot.

Foulger Pratt built this awareness directly into their underwriting on Accolade. Meroe acknowledged that 243 luxury units "felt a little heavy" for downtown Washington to absorb at the time. That is not a small admission from a developer who went ahead and built them anyway — it is a candid statement that even well-executed conventional conversion programs carry a demand absorption risk that requires hedging. The hedge, as we will examine, is the product mix.

You are also converting a building in a location where residential demand is structurally weak. Central business districts in most American cities lack the grocery stores, schools, and neighborhood fabric that drive long-term residential demand. The same CBD adjacency that made the building valuable as an office is a liability for commodity rental housing.

Converting an office to residential is a commoditization play. The building's best asset — its location — becomes its biggest liability.

The Brookings analysis put it plainly: most conversion projects remain financially challenging in downtowns across the country. Policy incentives — tax abatements, historic tax credits, low-interest loans — are not sweeteners. They are the mechanism by which most of these projects pencil at all. When the incentive disappears, so does the thesis. Foulger Pratt's experience at Accolade confirms this from the field: they financed the project in 2022, before D.C.'s conversion abatement programs were in place, and had to make the numbers work without them.

IV. RESIDENTIAL VERSUS RESIDENCES

Here is where the CRED analysis diverges from the market consensus. The office-to-residential conversion debate is being conducted almost entirely within a single product type — standard multifamily rental housing. The more important question is whether there is a different product the building is better suited to produce.

The answer is yes. And the difference is not cosmetic.

Residences — whether branded, serviced, or structured around flexible tenancy — use the same physical inputs but produce a categorically different financial output. The features that destroy value in a standard residential conversion become neutral or advantageous in a residences product.

The deep floor plate that kills the standard apartment layout — the 45–60 feet beyond the window line that produces dark, narrow units — becomes home offices, private gyms, chef's kitchens, and generous storage in a residences product. Large-format floor areas, which require expensive light wells in residential conversion, support oversized unit configurations that are genuinely desirable at the high end. The same geometry that is a problem for a $2,500-per-month studio is a feature for a $12,000-per-month two-bedroom residence.

Meroe's own conclusion at Accolade speaks directly to this logic. "We need to be looking at these buildings as platforms for multiple uses and multiple different experiences," he said. "It doesn't have to be another form of housing. If there's a couple of floors you want to peel off and you want to make it something else, whatever it is that you think can fit the space, I encourage people to do it." This is not an academic recommendation. It is the practitioner's lesson from a completed project.

The revenue numbers reflect this. Branded or premium residences typically command 20–40% higher prices per square foot than comparable non-branded luxury units. Rental rates run 15–25% higher than equivalent non-branded equivalents. The construction cost premium for a residences finish over a standard residential conversion is approximately 15–30%. The revenue premium is 100–200%.

The CBD location — which is a liability for commodity housing — is an asset for residences. The business traveler, the relocating executive, the international buyer seeking a pied-à-terre: these users pay a premium precisely because the building is adjacent to their professional corridor. The address that depresses long-term residential demand enhances residences demand.

This is not a positioning argument. It is a structural argument about which product the building is actually built for.

V. THE MARKET DEPTH OBJECTION — AND WHY IT HAS AN ANSWER

The legitimate pushback on the residences thesis has always been demand depth. The pool of buyers and tenants for premium branded residences is genuinely thinner than for standard rental apartments. In an oversupplied luxury market, a residences product without differentiated programming is just an expensive apartment with a different name on the door.

This is the right objection to raise. It is also the objection that Foulger Pratt answered empirically at Accolade — not in theory, but in executed square footage. Of the building's 243 units, 57 were purpose-built as short-term-stay units that Meroe described as operating "like a furnished hotel." The developer's own words: introducing that use "was very accretive" and gave "the building a better chance to be successful." The mixed-tenure structure was not an afterthought. It was the mechanism by which an otherwise heavy absorption challenge became manageable.

Accolade demonstrates what the data predicts. Two mechanisms work together to resolve the thin-market problem.

The first is the flex lease. A 90-day minimum tenancy — calibrated to land cleanly in the residential band under most U.S. regulatory frameworks — opens access to the corporate housing and serviced apartment market. That market was valued at $13.8 billion in 2024 and is projected to reach $44 billion by 2033, growing at a 14.5% compound annual rate. The average corporate housing stay is approximately 83 days. The product is not competing against luxury apartment oversupply. It is competing against a segment that is undersupplied, premium-priced, and growing at nearly three times the rate of the broader rental market.

The revenue math on flex leasing is direct. A 90-day corporate tenancy at $150 per night generates $13,500 in gross revenue. A standard 12-month lease on the same unit at $3,500 per month generates $42,000 per year. Cycling the flex unit four times annually produces $54,000 — a 28% revenue premium on identical physical space, at no additional construction cost. In markets where the flex product can operate year-round, the premium compounds further.

The flex lease also solves the CBD demand problem. The demand that naturally exists in a business corridor — corporate travelers, executives in transition, project teams, consulting deployments — is precisely the demand a flex product captures. The building is not fighting against its location. It is monetizing it. Meroe recognized this at Accolade. The furnished, hotel-adjacent units were not a concession to slow apartment absorption. They were the product best suited to the building's address.

The second mechanism is real estate programming. CRED REP builds a curated experience layer into the building that generates demand the physical asset cannot produce on its own. A curated speakers series, a chef residency, a wellness activation, a business network — this programming creates foot traffic from non-residents, and converts a fraction of that traffic into future tenants or flex occupants. It also generates ancillary revenue that does not exist in standard multifamily operations: on-site services — personal chefs, wellness programming, concierge functions — have demonstrated the capacity to add $50,000–$150,000 annually in NOI across a 150-unit building, entirely independent of unit occupancy.

The demand stack that results is not a single thin pool. It is layered: long-term premium residents anchoring the building, corporate flex tenants cycling through the high-demand corporate calendar, and program-driven traffic generating ancillary income that compounds the base. The building is no longer a residential asset competing in an oversupplied rental market. It is a living product competing in a serviced accommodations market that is growing fast and undersupplied in the locations office buildings occupy.

VI. THE OBBBA TAX LAYER: BENEFIT FOR BOTH, RISK FOR ONE

The One Big Beautiful Bill Act, signed July 4, 2025, permanently reinstated 100% bonus depreciation for qualifying property acquired after January 19, 2025. For office conversion developers, this is a meaningful cash flow event. On a $20M renovation, a cost segregation study typically reclassifies 20–35% of the building basis — FF&E, specialized electrical, plumbing fixtures, flooring, cabinetry — into 5-, 7-, and 15-year property eligible for full first-year expensing. At a 37% tax rate, that is $1.5M–$2.6M in immediate tax savings that was not reliably available before the legislation. The OBBBA also permanently preserves Qualified Improvement Property at a 15-year recovery period with 100% bonus depreciation eligibility — meaning interior conversion work on a former commercial building qualifies for the same accelerated treatment.

Both strategies — residential and residences — can access this benefit. The legislation does not distinguish between them. But it creates a new asymmetry that most developers are not modeling.

The recapture structure under the OBBBA is bifurcated and steep. Personal property components reclassified through cost segregation face ordinary income recapture at rates up to 37% on sale — not the 25% cap that applies to structural Section 1250 real property. Any bonus depreciation claimed on QIP is also recaptured at ordinary income rates. A developer who takes a $5M first-year deduction and then exits the asset in year three is not executing a deferral strategy. They are facing a recapture bill at ordinary income rates on the precise assets they wrote off — at the precise moment they are trying to close a sale.

The OBBBA benefit is real. But it is structured to reward long holds — and the standard residential conversion is a product with elevated exit risk in an oversupplied market.

For the standard residential conversion, this correlation is dangerous. The product — commodity rental units in CBD locations with thin demand — is the most likely candidate in the current market for a distressed or compressed-timeline exit. The developer who took the largest first-year deduction is also the developer most exposed to a forced sale at ordinary income recapture rates. The incentive and the risk are now traveling in the same direction.

The Accolade experience adds a further real-world note here. Foulger Pratt financed the project in 2022, before D.C.'s conversion abatement programs existed, and completed it without them. For developers who began projects expecting incentives that arrived late or not at all, the OBBBA's permanent bonus depreciation is a meaningful retroactive tailwind — but only for assets acquired after January 19, 2025. Projects underwritten before that date, which represents a significant share of the current conversion pipeline, do not qualify for the full benefit.

For the residences product, the dynamic inverts. The higher FF&E specification of a residences build — custom millwork, smart systems, premium kitchen packages, lighting design — shifts a greater proportion of the renovation budget into the personal property and QIP categories that qualify for bonus depreciation. The Year 1 benefit is at least as large, and likely larger, than for a standard residential conversion. But the residences operating model — long-term anchor tenants, flex lease cycling, programming-driven NOI — is structured for multi-year hold. The recapture liability has time to extinguish naturally. Five-year personal property components reach zero basis in five years. The developer who holds does not owe recapture because there is nothing left to recapture.

One further consideration: most states do not conform to federal bonus depreciation rules. California, New York, and New Jersey — three of the most active office conversion markets — either decouple entirely or provide only partial conformity. Developers underwriting the full federal benefit without modeling state-level addbacks are systematically overestimating Year 1 net benefit in the markets where conversions are most concentrated.

The OBBBA does not change the thesis. It sharpens it. The tax benefit is real for both strategies. The recapture risk falls asymmetrically on the strategy with the weaker NOI, the thinner demand, and the greater likelihood of an early exit. That strategy is standard residential.

VII. WHAT THIS DOESN'T SOLVE

This analysis would be incomplete without acknowledging what the residences model does not resolve.

Regulatory exposure is real. The 90-day flex lease sits in a legal band that varies by jurisdiction. New York City's short-term rental restrictions are aggressive; some municipalities have extended limitations to leases under 180 days. The flex product requires specific legal structuring in each target market, not a blanket assumption of availability. Operators who treat the regulatory framework as a formality will find it becomes an existential problem — a lesson Meroe's team absorbed firsthand in navigating D.C.'s layered code requirements.

Operating costs are heavier. A managed residences building with programming and flex cycling has more moving parts than a standard multifamily asset. Turnover costs are higher, staffing is more complex, and if the building participates in a managed rental pool, management fees can absorb 25–50% of gross revenue. The NOI premium is real; so is the margin compression from a poorly structured operating agreement.

Brand licensing has real costs. For developers new to the hospitality-adjacent product, the design consultation requirements, FF&E packages, and royalty structures associated with a true branded residences product are frequently underwritten optimistically. The brand is a trust mark that commands the premium. It is also a cost center that requires careful negotiation.

Capital availability is constrained in ways that have nothing to do with the project. Meroe was candid: even when a second conversion looks viable, convincing equity partners to commit to a specific downtown market — particularly in a period of political and economic uncertainty — is a separate challenge from underwriting the deal. Good projects are being shelved not because the numbers don't work but because the money is waiting.

And market timing matters. The corporate housing market is growing, but it is not homogeneous. A flex residences product in a market without a robust corporate demand base — no major employer relocations, no project-driven influx, weak convention economy — will struggle to cycle units at the rates that justify the model. The thesis is strongest in markets where the CBD is genuinely active, not merely geographically central.

Every viable real estate strategy has failure modes. The residences model solves most of the problems the residential model cannot. It does not solve all of them, and nothing does. The question is whether the risk-adjusted return profile is superior. The data says it is. So does Accolade.

VIII. THE LINE THAT MATTERS

The office vacancy crisis produced a genuine opportunity. It also produced a gold rush, and gold rushes have a consistent track record of enriching the few who understood the asset and impoverishing the many who followed the crowd.

The developers getting hurt are those who bought cheap, assumed incentive programs would close the gap, and produced commodity rental units in locations that do not support commodity rental demand. They are following each other into the same trade, in the same markets, with the same product. The supply glut they are creating will take years to absorb. Some of them will not survive the lease-up.