Adaptive reuse pro forma: when conversion cost exceeds finished value

An analyst at an institutional real estate investor asks Compound whether an office-to-apartment conversion is worth what it costs. Compound picks a 1962 Hartford office complex from CMBS filings and builds the model from Schedule III, a City of Denver conversion case study, BLS cost indexes, Zillow, Apartment List, HUD rents, Connecticut sale records and the SOFR forward curve.

Compound Research · Updated Nov 15, 2026 · 14 min read

Short answer

An adaptive reuse pro forma tests whether the finished value of a converted building carries the cost of converting it. Compound's model of a Hartford office complex from CMBS filings, at a Denver case study's cost per square foot, shows 659 apartments costing $383.1 million against $128.0 million of finished value.

Try it now
On this page (10)
  1. 1. What goes into an adaptive reuse pro forma
  2. 2. The existing shell: the building and the purchase basis
  3. 3. Conversion cost per square foot: three hard-cost groups
  4. 4. Entitlement carry, the draw curve and capitalized interest
  5. 5. Apartment rents and finished value per unit
  6. 6. Does the finished value carry the cost? Two exits side by side
  7. 7. Entitlement delay sensitivity: 6, 12 and 18 months
  8. 8. What public data cannot tell you about a conversion candidate
  9. 9. Our view: start with value per unit against cost per unit
  10. 10. Frequently asked questions

Converting this office complex into 659 apartments costs $383.1 million, and the finished apartments are worth $128.0 million. That is the base case of this model for Constitution Plaza, a 659,315 square foot office property in Hartford, Connecticut, as its loan’s CMBS filings describe it. The purchase basis is $13.0 million, or $19.72 per square foot, and it is 3.4% of the budget. Nearly all of the cost is the work.

Read the cost with its source in mind. The model prices the work from the construction budget in a City of Denver conversion case study and applies it to every square foot, which gives $581,398 per apartment. Two completed Hartford conversions recorded in the workbook cost about $137,000 and $296,000 per unit (our arithmetic), so treat the $383.1 million as the result of these inputs.

An adaptive reuse pro forma answers one question first: is the finished building worth more than the purchase price plus the cost of converting it? We asked Compound to build that model on a real building from public data. It ran on November 15, 2026 and returned one Excel workbook and a summary.

The building’s size, age, occupancy, income, appraisal and loan come from public filings. The purchase price is an assumption set by the request: the lower of the latest appraised value and the outstanding loan balance. Nothing in the filings says the property can be bought at that figure. The shortfall is the output of the model’s inputs. It is neither a forecast for this property nor advice to buy or sell it.

What the workbook shows:

  • Hard cost alone is more than the finished value. Three hard-cost groups total $350.03 per square foot at today’s prices, or $230.8 million.
  • Cost per unit is $581,398 and value per unit is $194,167. The cost comes from a Denver building’s budget, and the value comes from Zillow home values by bedroom for the Hartford metro.
  • Both exits lose money. Selling at completion loses $257.7 million of equity. Holding to stabilization loses $291.1 million, a levered IRR of negative 26.0%. The two sale prices rest on different evidence: home values at completion and a cap rate at the end of the hold. The building comes from CMBS loan-level filings on SEC EDGAR. The basis comparison comes from REIT Schedule III data. Both loans float over Compound’s SOFR forward curve, and the permanent rate starts from the Daily Treasury Par Yield Curve.
Prompt to try

“I am weighing an office-to-apartment conversion and want an adaptive reuse pro forma that shows whether the finished value carries the cost, anchored on a real building. From the CMBS loan-level filings, pick one downtown office property in a Midwest or Northeast city, built before 1990, that is less than 60% occupied or whose loan is past its maturity date, and whose latest report gives net rentable area, year built, occupancy, net operating income, appraised value and the loan balance. Use it as the conversion candidate and set the purchase basis at the lower of its latest appraised value and the outstanding loan balance; compare that basis per square foot with what public office REITs carry buildings at in the same state (Schedule III). Set the unit count from the rentable area at an assumed efficiency and average unit size. Set hard costs in three groups (keep and repair the structure, envelope and systems; selective demolition; residential build-out) as analyst assumptions per square foot from published conversion cost ranges, brought to today with the BLS construction cost PPI and construction wage series. Set apartment rents from Zillow ZORI and Apartment List for the metro and HUD Fair Market Rents for the ZIP code, and the sale value per unit from comparable sales, using Zillow ZHVI for the metro and county property sale records where the state is covered, with the exit cap rate as a cross-check. Carry property tax, insurance and acquisition loan interest through an entitlement period, then a 24-month construction period with a monthly draw curve. Price the loans off the SOFR forward curve plus a spread, with interest accrued on the running balance and capitalized, and an interest reserve sized at closing and shown as a use. Build five sections: existing shell, entitlement, construction, exit, and sources and uses. Report total project cost with capitalized interest and reserve, cost per unit, and two sets of returns side by side: sell at completion, and hold to stabilization with a permanent loan off the 10-year Treasury yield. Add a sensitivity of profit and levered IRR to entitlement delays of 6, 12 and 18 months. Flag a finished value below total cost and an interest reserve that runs out before completion. Use the stored public data wherever it has the figure, and web search only for what it lacks. Put every input on an Assumptions sheet with one column that says where it came from: a cited source, or "analyst assumption" for a figure I would set myself, with the reason for the value you chose. Every output must be a live formula that traces back to the Assumptions sheet. Add a Sources sheet listing each data source, what it gave and its date. Deliver one Excel workbook and a short summary of the result, the flags, and what I would need from the seller to firm it up. Cite every sourced figure.”

Try this prompt free →

That is the request as we sent it, without its closing line, which asks that no name or email be put in the workbook.

What goes into an adaptive reuse pro forma

The model has five sections: the existing shell, entitlement, construction, exit, and sources and uses. Filings and public series supply the building, the cost indexes, the rents, the home values and the rates, a City of Denver case study supplies the hard costs, and the unit mix and the loan terms are analyst assumptions.

The model assumes vacant possession when construction starts. It keeps 0% of the office net operating income during entitlement and does not model the cost of buying out leases. The timeline is 12 months of entitlement, 24 months of construction, then lease-up.

Each input on the Assumptions sheet carries a citation, the word “Request”, or the words “analyst assumption” with a reason. The 75% residential efficiency, for example, is the average of the Denver case study and a New York City Comptroller note.

Two sets of figures came from the web: Connecticut’s recorded sales from the state’s open-data portal, and Apartment List’s rents by bedroom from its listing pages. The Sources sheet lists 26 sources.

The existing shell: the building and the purchase basis

Compound picked Constitution Plaza in Hartford from the CMBS loan-level filings and set the basis at its $13,000,000 appraisal. That is $19.72 per square foot, 6.8% of the gross value at which a public REIT carried two Connecticut office buildings.

The filing for Wells Fargo Commercial Mortgage Trust 2018-C46, for the report period ending October 13, 2026, lists the property as 1, 10, 100, 248, 250 & 260 Constitution Plaza, Hartford, CT 06103. It was built in 1962 and last renovated in 2004. Physical occupancy is 73%, above the request’s 60% line, so the building qualifies on the other test. The loan’s maturity date was May 6, 2023, which is 43 months before the model’s closing date. The filing reports the loan as not repaid at that date.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
C20fx=MIN(C16,C19)
BCDEFGH
5CANDIDATE PROPERTYValue
6PropertyConstitution Plaza (1, 10, 100, 248, 250 & 260 Constitution Plaza), Hartford, CT 06103
7Net rentable area (sf)659,315
8Year built1962
9Physical occupancy73%
10Revenue, trailing 12 months ($)14,078,747
11Operating expenses, trailing 12 months ($)11,798,176
12Net operating income, trailing 12 months ($)2,280,571
13NOI per sf ($)3.46
15PURCHASE BASIS
16Latest appraised value ($)13,000,000
17Loan balance, WFCM 2018-C46 note ($)16,369,863
18Loan balance, JPMDB 2018-C8 note ($)24,554,793
19Total outstanding loan balance ($)40,924,656
20Purchase basis: lower of appraisal and loan balance ($)13,000,000
21Purchase basis per sf ($)19.72
22In-place NOI yield on basis17.5%
23Loan maturity date6-May-2023
24Months past maturity at closing43
25Qualifies: built before 1990 and (occupancy below 60% or loan past maturity)Yes
27SCHEDULE III COMPARISON: PUBLIC OFFICE REIT CARRYING VALUES IN CONNECTICUT
28Gross carrying ($)Accum. depreciation ($)Net carrying ($)Rentable sfGross $/sfNet $/sf
29Metro Center, Stamford (ESRT, FY2024)76,551,00045,505,00031,046,000282,276271.19109.98
30First Stamford Place, Stamford (ESRT, FY2023)232,184,000113,916,000118,268,000781,731297.01151.29
31Average of the two REIT buildings308,735,000159,421,000149,314,0001,064,007290.16140.33
32Constitution Plaza purchase basis13,000,00013,000,000659,31519.7219.72
33Basis as a share of REIT carrying value per sf6.8%14.1%
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The Existing Shell sheet, columns B to H, as Compound wrote it. We leave out the sheet title and the note under the Schedule III block. The formula bar shows the basis rule from the request: the lower of the appraisal and the total loan balance.

For the 12 months to September 30, 2025 the filing shows net operating income of $2,280,571, or $3.46 per square foot. The latest MAI appraisal is $13,000,000, dated May 29, 2025. The loan is two pari passu notes, one in the 2018-C46 trust and one in JPMDB Commercial Mortgage Securities Trust 2018-C8. Together their balance is $40,924,656, about 3.1 times the appraisal (our arithmetic). The basis cell is =MIN(C16,C19), and the appraisal is the smaller figure.

At that basis the in-place income is a 17.5% yield, and the model does not count on it. The building is 73% occupied in the filing, and the model assumes it is empty when the work starts. What it would cost to reach that point is not in the model.

The request also asked how the basis compares with what public office REITs carry buildings at in the same state. The Schedule III data has two Connecticut office lines, both from Empire State Realty Trust and both in Stamford. The two together average $290.16 gross and $140.33 net of depreciation.

The $19.72 basis is 6.8% of the gross figure and 14.1% of the net figure. Read that comparison narrowly: it covers two buildings in another city, and the workbook notes that both have since left the REIT’s portfolio.

Conversion cost per square foot: three hard-cost groups

Hard cost is $322.45 per square foot at April 2025 prices, taken from a City of Denver conversion case study, and $350.03 today after escalation with two BLS series. With contingency, soft costs, the developer fee and inflation to the draw dates, the development budget is $351,946,923 before land, carry and financing.

The unit count comes from area. At 75% efficiency the 659,315 square feet leave 494,486 square feet of net residential area, and at an average of 750 square feet the result is 659 units: 198 studios, 330 one-bedrooms and 131 two-bedrooms. The unit size and mix are analyst assumptions.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
C16fx=1+Assumptions!$C$46*C14+(1-Assumptions!$C$46)*C15
BCDEF
13HARD COST ESCALATION, APR-2025 TO SEP-2026
14BLS PPI, inputs to multifamily construction: change11.2%
15BLS avg hourly earnings, construction of buildings: change5.9%
16Blended escalation factor1.086
18DEVELOPMENT BUDGET$/sf, Apr-2025$/sf, todayTotal ($)$/unit
19Keep and repair structure, envelope and systems157.74171.23112,893,954171,311
20Selective demolition and abatement12.3313.398,825,61113,392
21Residential build-out152.38165.42109,063,209165,498
22Total hard cost ($)322.45350.03230,782,774350,201
23Hard cost contingency52.5134,617,41652,530
24Soft costs70.0146,156,55570,040
25Developer fee18.9012,462,27018,911
26Hard cost escalation to the planned draw dates42.3627,927,90842,379
27Development budget before land, carry and financing ($)533.81351,946,923534,062
28Check: total hard cost $/sf, today, against the CBRE $250-$650 rangeInside range
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The escalation block and the development budget on the Construction sheet, columns B to F, as Compound wrote them. We leave out the unit program above and the reserve and draw-curve blocks below. The formula bar shows the blended escalation factor: half the change in the BLS materials index and half the change in construction wages.

The request set three hard-cost groups. The workbook fills them from the construction budget in the City and County of Denver’s April 2025 case study of an office conversion with 158,000 square feet of building area. Each line of that budget is divided by the area and marked up 10.4% for the general contractor’s costs and builder’s risk.

Keeping and repairing the structure, envelope and systems is $157.74 per square foot. Selective demolition and abatement is $12.33. Residential build-out of units and common areas is $152.38. The cell note says these are Denver prices with no adjustment for Hartford.

Two BLS series bring those April 2025 prices to today. The producer price index for inputs to multifamily construction rose 11.2% between April 2025 and September 2026, and average hourly earnings in construction of buildings rose 5.9%. The model weights the two at 50% each, and the factor is 1.086.

At today’s prices the three groups are $171.23, $13.39 and $165.42 per square foot. Total hard cost is $230,782,774, or $350,201 per unit. A check cell compares the $350.03 with a published range of $250 to $650 per square foot from CBRE and reads “Inside range”. The cell note on demolition and abatement says a 1962 building may need more than the Denver allowance.

Four lines sit under hard cost: a 15% contingency (the case study’s 10% design and 5% construction contingencies), 20% soft costs from a Citizens Budget Commission report, a 4% developer fee, and $27,927,908 of inflation at 5% a year to each draw date. The inflation formula covers hard cost, contingency and the developer fee. The budget is $533.81 per square foot and $534,062 per unit.

Entitlement carry, the draw curve and capitalized interest

The model holds the shell for 12 months of entitlement and 24 months of construction, and prices both loans off the SOFR forward curve. Tax and insurance carry is $3,143,615, origination fees are $2,190,586 and the interest reserve sized at closing is $16,465,782.

During the entitlement year the model pays tax and insurance on a building that earns nothing. Connecticut’s 70% assessment ratio and a combined Hartford rate of 71.6794 mills make the tax $652,283 a year, and insurance at $0.60 per square foot is $395,589. Together they are $87,323 a month.

An acquisition loan funds 50% of the basis, or $6,500,000. Its rate is the average 1-month forward SOFR over the entitlement period, 4.52%, plus a 4.00% spread, or 8.52%. With 35% of soft costs spent on design and permitting, entitlement-period cost is $17,737,472.

Construction draws follow a 24-month S-curve. Equity goes in first. The sponsor’s $167,948,328 funds costs before the construction loan does, and the loan balance reaches $215,192,926 at completion in month 36. The construction loan prices at forward SOFR plus 3.75%, and the average over the construction period is 4.68%, or 8.43% all in.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
C35fx=(C32+C33+C34)*(1+Assumptions!$C$124)
BCDEF
5USES$$/sf$/unit% of total
6Purchase basis13,000,0002019,7273.4%
7Acquisition closing costs260,00003950.1%
8Hard cost: structure, envelope and systems112,893,954171171,31129.2%
9Hard cost: selective demolition and abatement8,825,6111313,3922.3%
10Hard cost: residential build-out109,063,209165165,49828.2%
11Hard cost escalation to draw dates (planned)27,927,9084242,3797.2%
12Hard cost contingency34,617,4165352,5308.9%
13Soft costs46,156,5557070,04011.9%
14Developer fee12,462,2701918,9113.2%
15Tax and insurance carry, planned entitlement + construction3,143,61554,7700.8%
16Costs before financing ($)368,350,537559558,95495.2%
17Loan origination fees2,190,58633,3240.6%
18Interest reserve16,465,7822524,9864.3%
19Total uses ($)387,006,905587587,264100.0%
21SOURCES
22Acquisition loan6,500,000109,8631.7%
23Construction loan commitment (incl. interest reserve)212,558,577322322,54754.9%
24Sponsor equity167,948,328255254,85343.4%
25Total sources ($)387,006,905587587,264100.0%
26Check: sources less uses0
28LOAN SIZING
29Debt for costs: loan-to-cost x costs before financing ($)202,592,796
30Construction-period average SOFR (planned)4.68%
31Construction loan all-in rate8.43%
32Reserve estimate: acquisition loan interest, entitlement ($)553,919
33Reserve estimate: interest on the rolled acquisition loan, construction ($)1,189,388
34Reserve estimate: interest on construction draws ($)13,225,585
35Interest reserve with cushion ($)16,465,782
36Total loan commitment ($)219,058,577
37Loan-to-cost on total uses56.6%
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The Sources & Uses sheet, columns B to F, as Compound wrote it: uses, sources and loan sizing. The formula bar shows the interest reserve: three interest estimates plus a cushion.

The sheet builds the interest reserve from three interest estimates, the largest being $13,225,585 on construction draws, and adds a 10% cushion to reach $16,465,782.

Total uses are $387,006,905, or $587,264 per unit. The loan is 55% of costs before financing plus the reserve, a total commitment of $219,058,577 and 56.6% of total uses. Sponsor equity is $167,948,328, or 43.4%.

Uses of funds at closing

Dollars. The three hard-cost groups are highlighted. The purchase basis is the first bar

Purchase basis$13.0MAcquisition closing costs$0.3MHard cost: structure, envelope and systems$112.9MHard cost: selective demolition and abatement$8.8MHard cost: residential build-out$109.1MHard cost escalation to draw dates$27.9MHard cost contingency$34.6MSoft costs$46.2MDeveloper fee$12.5MTax and insurance carry$3.1MLoan origination fees$2.2MInterest reserve$16.5M

Compound workbook, Sources & Uses sheet, cells C6 to C18. Hard costs are analyst assumptions per square foot, escalated with BLS series.

The three hard-cost groups are 59.6% of uses (our arithmetic). The purchase basis is 3.4%. Total project cost less the basis and its closing costs is about $369.9 million (our arithmetic). A free building would still cost more than twice the finished value to convert.

The monthly schedule then tests the reserve. It capitalizes $12,600,130 of interest through completion, so $3,865,651 of the reserve is unused. Total project cost with capitalized interest is $383,141,253.

Apartment rents and finished value per unit

Underwritten rent is $1,778 a month across the unit mix, taken from Apartment List’s Downtown Hartford averages with no premium. Finished value is $194,167 per unit at completion, from Zillow home values by bedroom, and all but one of the workbook’s reference values are lower.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
F18fx=SUMPRODUCT(C18:E18,$C$10:$E$10)
BCDEF
5UNDERWRITTEN RENTS (TODAY'S $ PER UNIT PER MONTH)StudioOne-bedTwo-bedWeighted
6HUD Small Area FMR, ZIP 06103 (FY2027)1,6501,8902,3701,914
7HUD FMR, Hartford metro (FY2027)1,3461,5431,9331,562
8Apartment List, Downtown Hartford (studio: on-plaza comp)1,5541,7402,2101,778
9Apartment List, Hartford city1,4321,5011,7171,524
10Unit mix30%50%20%100%
11Underwritten rent1,5541,7402,2101,778
12Zillow ZORI, ZIP 06103, all units1,579
13Zillow ZORI, Hartford metro, multifamily1,911
14Underwritten rent vs ZORI ZIP 0610312.6%
15Underwritten rent vs HUD SAFMR(7.1%)
17SALE VALUE PER UNIT (COMPARABLE SALES, TODAY'S $)StudioOne-bedTwo-bedWeighted
18Zillow ZHVI by bedroom, Hartford metro138,424162,852300,072182,968
19Adjustment to comp value0.0%
20Comp value per unit, today ($)182,968
21CT recorded sales: Hartford condos, mean151,215
22CT recorded sales: Hartford condos, median108,000
23Hartford County multifamily, 2025 price per unit117,031
24Zillow ZHVI, metro condos (all sizes)269,064
25Comp value vs CT recorded-sales mean21.0%
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The rent and sale-value blocks of the Exit sheet, columns B to F, as Compound wrote them. The formula bar shows the weighted sale value per unit: the Zillow value for each bedroom count times the unit mix.

Apartment List’s Downtown Hartford averages are $1,740 and $2,210 for one and two bedrooms, and the studio rent of $1,554 is a listing at 5 Constitution Plaza.

Monthly rent per unit by source, weighted by the unit mix

Dollars per unit per month. The underwritten rent is highlighted. ZORI figures are single series with no bedroom detail

HUD Small Area FMR, ZIP 06103, FY2027$1,914Zillow ZORI, Hartford metro, multifamily$1,911Underwritten: ApartmentList, Downtown Hartford$1,778Zillow ZORI, ZIP 06103, all units$1,579HUD FMR, Hartford metro, FY2027$1,562Apartment List, Hartford city$1,524

Compound workbook, Exit sheet, cells F6 to F13. HUD figures are FY2027, Zillow ZORI is September 2026, and Apartment List pages were read on November 15, 2026.

Compound underwrote the downtown averages, $1,778 weighted, with a 0% premium for new units. That is 12.6% above the ZORI figure for the ZIP and 7.1% below the HUD figure.

For value, the request asked for Zillow’s ZHVI for the metro. The September 2026 values for the Hartford metro are $162,852 for a one-bedroom home and $300,072 for a two-bedroom. ZHVI has no studio series, so the model sets a studio at 85% of the one-bedroom value, $138,424. Weighted by the mix, the value is $182,968 per unit today. Growth of 2% a year to completion makes it $194,167, and 659 units make a finished value of $127,955,895.

The sheet then lists what else the market says. Connecticut’s recorded sales show 178 Hartford condo sales from October 2024 to September 2025 with a mean of $151,215 and a median of $108,000. A broker’s county market report puts 2025 multifamily sales in Hartford County at $117,031 per unit and an 8.3% average cap rate. The one higher reference is ZHVI for metro condos of all sizes, $269,064.

Cost per unit against every value per unit in the workbook

Dollars per apartment. Total project cost is the first bar. The value the model uses is highlighted

Total project cost per unit$581,398Comp value per unit at completion:Zillow ZHVI blend, grown 2% a year$194,167Zillow ZHVI blend by bedroom,Hartford metro, today$182,968Value per unit at the 7.25%exit cap, first stabilized year$163,670Connecticut recorded sales,Hartford condos, mean$151,215Hartford County multifamily sales, 2025, per unit$117,031Connecticut recorded sales,Hartford condos, median$108,000

Compound workbook, Exit sheet, cells C44, C39, F20, C32, F21, F23 and F22. Zillow ZHVI is September 2026. Connecticut recorded sales are from the state's open-data portal.

The exit cap rate is the cross-check the request asked for. In the first stabilized year the model’s net operating income is $7,819,721, or $11,866 per unit. At a 7.25% exit cap, an analyst assumption, the income is worth $107,858,223, or $163,670 per unit. That is 19.1% below the comp value at the same date.

ZHVI measures for-sale homes across a metro, and the subject is rental apartments in one downtown. None of the references is a sale of a newly converted downtown apartment building. We would treat $194,167 as the generous end of the evidence in this workbook.

Does the finished value carry the cost? Two exits side by side

It does not. Total project cost with capitalized interest is $383,141,253 and finished value at completion is $127,955,895, so the model raises its flag and both exits lose money.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
C60fx=IFERROR(XIRR('Monthly CF'!$E$91:$FP$91,'Monthly CF'!$E$23:$FP$23,-0.1),"n.m.")
BCD
36FINISHED VALUE AGAINST TOTAL COST (AT COMPLETION)
37Completion dateNov-2029
38Apartment units659
39Comp value per unit at completion ($)194,167
40Project costs before interest, through completion ($)370,541,123
41Interest capitalized through completion ($)12,600,130
42Total project cost incl. capitalized interest ($)383,141,253
43Budgeted total uses incl. interest reserve ($)387,006,905
44Total project cost per unit ($)581,398
45Total project cost per sf ($)581
46Finished value at completion ($)127,955,895
47Finished value less total project cost ($)(255,185,359)
48Flag: finished value below total costFLAG: finished value $128.0M is below total cost $383.1M
49Flag: interest reserve runs out before completionOK: reserve covers interest to completion, $3.9M left
51RETURNSSell at completionHold to stabilization
52Exit dateNov-2029Dec-2036
53Gross sale price ($)127,955,895119,704,561
54Net sale proceeds ($)125,396,777117,310,470
55Loan repaid at exit ($)215,192,92670,993,580
56Equity invested ($)257,744,477348,336,452
57Equity distributions ($)057,266,747
58Levered profit ($)(257,744,477)(291,069,705)
59Equity multiple0.00x0.16x
60Levered IRR (annualized)n.m.(26.0%)
61Unlevered profit ($)(245,144,346)(212,124,860)
62Unlevered IRR (annualized)(70.0%)(10.5%)
63Permanent loan at stabilization ($)75,500,756
64Permanent loan rate6.90%
65DSCR on forward NOI at stabilization1.31x
66Yield on total project cost (forward NOI at stabilization)2.0%
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The value-against-cost block and the returns block of the Exit sheet, columns B to D, as Compound wrote them. The formula bar shows the levered IRR of the sale case: an XIRR on the monthly equity cash flows that returns n.m. when no cash flow is positive.

The flag reads “FLAG: finished value $128.0M is below total cost $383.1M”. Finished value is 33.4% of cost, and cost per unit is 2.99x value per unit (our arithmetic). The Summary sheet states the same gap as a break-even: each apartment would need to be worth $581,398.

Sell at completion. The sale in November 2029 brings $127,955,895, or $125,396,777 after 2% selling costs. That price is the comp value from Zillow home values, applied to a building that has not yet leased. The loan balance is $215,192,926. Sale proceeds fall $89.8 million short of the loan (our arithmetic), and the sponsor pays that on top of the equity commitment. Equity invested is $257,744,477, distributions are zero and the equity multiple is 0.00x. The levered IRR cell reads “n.m.” and the unlevered IRR is negative 70.0%.

Hold to stabilization. The hold ends in December 2036. The permanent loan prices at the 10-year Treasury par yield of 5.15% plus an assumed 1.75%, or 6.90%. It is the smaller of 70% of value at the exit cap and the loan that income covers 1.25x. The result is $75,500,756 with coverage of 1.31x.

The monthly schedule shows the construction loan at $252,290,359 when the permanent loan funds. The permanent loan repays part of it, and the sponsor funds the other $176.8 million or so (our arithmetic). Equity invested rises to $348,336,452, distributions total $57,266,747, and the equity multiple is 0.16x. The levered IRR is negative 26.0% and the unlevered IRR is negative 10.5%.

Holding loses more dollars than selling, $291,069,705 against $257,744,477, for two reasons. First, the two exits use different value bases. The sale at completion uses the comp value, $127,955,895. The sale at the end of the hold capitalizes forward net operating income at the 7.25% exit cap, which gives $119,704,561 seven years later. In the first stabilized year the cap-rate method gives a value 19.1% below the comp value.

Second, carry. The building earns a 2.0% yield on its total cost while debt costs 6.90% or more. For value at the 7.25% cap to equal total cost, net operating income would need to be about $27.8 million, roughly 3.6 times the model’s figure (our arithmetic).

Entitlement delay sensitivity: 6, 12 and 18 months

Each six months of delay adds between $9.0 million and $9.5 million to total cost (our arithmetic on the table). The interest reserve lasts to completion in all three cases, with $928,752 left after an 18-month delay.

Constitution_Plaza_Office_to_Apartment_Pro_Forma.xlsx
C8fx=Exit!C58
BCDEFGHI
7Entitlement delay (months)Sell: levered profit ($)Sell: levered IRRHold: levered profit ($)Hold: levered IRRTotal project cost ($)Finished value less total cost ($)Reserve left at completion ($)
8Live case(257,744,477)n.m.(291,069,705)(26.0%)383,141,253(255,185,359)3,865,651
90(257,744,477)n.m.(291,069,705)(26.0%)383,141,253(255,185,359)3,865,651
106(265,494,933)n.m.(300,495,520)(26.3%)392,139,469(262,910,350)2,942,888
1112(273,510,879)n.m.(310,395,476)(26.7%)401,415,592(270,900,579)1,928,703
1218(281,691,143)n.m.(320,488,481)(27.1%)410,868,571(279,054,869)928,752
SummaryAssumptionsExisting ShellEntitlementConstructionExitSources & UsesSensitivityMonthly CFSources
The delay table on the Sensitivity sheet, columns B to I, as Compound wrote it. Row 8 links to the live case. Rows 9 to 12 are an Excel what-if data table that reruns the monthly schedule for each delay.

The delay is one input on the Assumptions sheet, and the whole model reruns for each row: the draw dates, the forward rates on those dates and the reserve balance. Total cost rises from $383,141,253 on plan to $410,868,571 at 18 months. The loss in the sale case grows from $257,744,477 to $281,691,143, and the levered IRR of the hold case moves from negative 26.0% to negative 27.1%.

Interest reserve left at completion, by entitlement delay

Dollars of the $16,465,782 reserve unused when construction ends

Plan, no delay$3.87M6-month delay$2.94M12-month delay$1.93M18-month delay$0.93M

Compound workbook, Sensitivity sheet, cells I9 to I12.

The reserve shrinks from $3,865,651 to $928,752 across the cases. In a conversion with a thin positive margin, a cost of this size for each six months could decide the outcome.

What public data cannot tell you about a conversion candidate

Public filings describe the building as a lender reports it: size, age, occupancy, income, appraisal and loan. A conversion depends on the building’s layout, its condition and its leases, and those come from the seller.

Compound’s summary ends with the list it would ask the seller for. Each item replaces a labeled assumption.

  • The rent roll, the leases and any termination rights. The model assumes vacant possession when construction starts and carries no buyout cost. The leases say when that can happen and what it costs.
  • Floor plans for each building. Floor-plate depth and the position of cores and windows set the efficiency and the unit count.
  • Condition and hazardous-materials surveys. These drive all three hard-cost groups, which are priced here from one Denver building’s budget.
  • The May 2025 appraisal and the terms of a sale. The basis here is a rule applied to a filed appraisal. The price a seller would accept is not in the filing.
  • The assessment, the tax bill, title and the plaza’s shared areas. The carry in the model applies the state’s 70% ratio and current mill rates to the basis, and what the six addresses share affects both cost and value. The $2,000 per unit of tax after completion assumes a tax agreement.

Our view: start with value per unit against cost per unit

A low basis does not make a conversion work. In this model the building could be free and conversion would still cost more than twice what the finished apartments are worth.

The basis is 3.4% of uses, and hard cost alone is more than the finished value. So the first test of a conversion is two numbers: all-in cost per unit and value per unit in that submarket. Here they are $581,398 and $194,167. The other side has weight, and it is local. The Sources sheet records two completed Hartford conversions from a Hartford Courant article of August 3, 2025: the plaza’s former hotel, converted to 190 units for $26M, and 777 Main, converted to 285 units for $84.5M. Those work out to about $137,000 and $296,000 per unit (our arithmetic), both far under this model’s $581,398. The workbook does not adjust those costs for date or say what they include.

The model’s cost comes from a different method: one Denver building’s budget per square foot, applied to all 659,315 square feet, or about 1,000 gross square feet for each apartment (our arithmetic). At the higher local figure, 659 units would cost about $195.4 million, still above the finished value. At the lower one they would cost about $90.2 million, below it (our arithmetic). We would not accept or reject a conversion on a cost per square foot borrowed from another city.

Three things would change our view: a survey and floor plans that cut hard cost per unit, public money such as tax credits or a tax agreement, and rent evidence from new downtown units that supports a premium over the average.

Other models answer the neighboring questions. A real estate development pro forma builds new apartments on land, where cost per square foot buys a floor plate designed for housing. An office lease rollover analysis keeps a building as office and asks what its leases are worth. A construction-to-permanent loan model isolates the takeout that the hold case here shows falling short.

Adaptive reuse pro forma (Excel)Compound’s workbook from this run: a Summary with the result, both flags and the delay cases; the Assumptions sheet with a source or a reason for every input and the SOFR forward curve; the Existing Shell, Entitlement, Construction, Exit and Sources & Uses sections; the Sensitivity data table; the Monthly CF engine; and the Sources list.Download · Free

Frequently asked questions

What is an adaptive reuse pro forma?

It is a model of converting an existing building to a new use: a purchase basis for the shell, a holding period for approvals, a construction budget and draw schedule, a finished value, and sources and uses. Its main output is whether the finished value is more than total cost.

How much does an office-to-apartment conversion cost per square foot?

Published ranges are wide. The workbook records a CBRE range of $250 to $650 per square foot. This model takes $322.45 of hard cost at April 2025 prices from a City of Denver case study, which becomes $350.03 today after BLS escalation. With contingency, soft costs, fee, inflation, carry and financing, total project cost is $581 per square foot.

What purchase price makes an office-to-residential conversion work?

Work backward: finished value, less conversion cost and financing, less a profit margin, is the most you can pay for the shell. In this model the answer is below zero. Total cost without the purchase is about $369.9 million (our arithmetic) against a finished value of $128.0 million. A price is an assumption here, set at the $13,000,000 appraisal.

How do entitlement delays affect a conversion pro forma?

A delay extends tax, insurance and loan interest on an empty building and moves every construction draw later at inflated prices. In this model each six months adds between $9.0 million and $9.5 million to total cost (our arithmetic).

How is an adaptive reuse pro forma different from a development pro forma?

A ground-up model starts with land. A conversion starts with a structure, so it adds demolition and abatement, repair of existing systems, a larger contingency and the loss of the old income. The existing floor plate sets how much of the area becomes rentable.

CoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix ResearchCoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix ResearchCoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix Research

See this workflow on your own deal

Try it now