Stuyvesant Town sold twice at nearly the same price — the capital stack decided which deal survived

The short answer to why Stuyvesant Town’s first sale failed: the buyers put in almost no money of their own, borrowed the rest against income that did not exist yet, and the income never arrived because a court said the law had never allowed it. Tishman Speyer and BlackRock Realty paid $5.4 billion in November 2006 on a $3 billion interest-only senior mortgage and roughly $1.4 billion of mezzanine debt stacked eleven levels deep.^5,6 Rental income covered forty per cent of the debt service from the first day.^8,9 The gap was paid out of a reserve fund, and when the reserve ran dry the deal was over.

Nine years later Blackstone and Ivanhoé Cambridge bought the same complex for $5.3 billion with about $2.6 billion of equity — roughly half the price, in cash, on the day.⁴³ The debt beside it was a single ten-year Fannie Mae loan that Fannie itself described as low-leverage.⁴⁷ That is the whole comparison, and it is why this brief’s central figure is two bars rather than a chart of rents.

Start with the seller, because the seller explains the timing

Metropolitan Life built Stuyvesant Town and Peter Cooper Village in the 1940s as a middle-class housing project on the cleared Gas House District, and held them for sixty years. What changed was not the buildings but the owner’s legal form. MetLife demutualised and went public in 2000, and a quiet, tax-advantaged, socially framed hold stopped making sense for a company that now had to report to shareholders every quarter.

On 17 October 2006 it announced the sale, and on 17 November it closed. The company’s own filing states the outcome plainly: a gain of approximately $3 billion, net of income taxes.^1,2 That figure is the least examined number in the whole story and arguably the most important one. Everything that followed — the litigation, the default, the five years of servicer control, the second sale — happened downstream of a seller who collected in full and left.

The auction was competitive and the record is unusually clear about it. Darcy Stacom of CB Richard Ellis ran the sale; roughly eight finalists reached the end. Apollo Real Estate Advisors, working with the Dermot Company, bid $5.33 billion. Related bid with Lehman Brothers; the Milstein brothers and Vornado bid. A tenant-organised bid led by City Council member Daniel Garodnick came in at $4.5 billion. Tishman Speyer and BlackRock signed at around ten in the morning on 17 October after an overnight negotiation, having gone about $70 million above the runner-up.^3,4

Seventy million dollars is a rounding error on $5.4 billion, and it is also the entire margin by which the losing bidders avoided the worst commercial real estate loss on record. Winning an auction and being right are different things.

What $5.4 billion actually bought

Eighty contiguous acres of Manhattan between First Avenue and the FDR Drive, 14th Street to 23rd Street, holding somewhere around eleven thousand two hundred apartments and about twenty-five thousand residents.³ The apartment count wobbles by source — 11,232 in the Times in 2006, 11,227 in 2010, 11,250 in two trade publications, 11,241 in Fannie Mae’s 2015 announcement — and no source explains the spread.^3,6,10,11,47

The building count does not wobble; it splits. The New York Times said 110 buildings in 2006 and again in 2010. Multifamily Executive, Commercial Property Executive and the Congressional Oversight Panel — the last citing SEC and Fitch filings directly — said 56.^3,5,10,11 Two well-sourced camps, no reconciliation, and it is worth noticing that the most heavily reported real estate transaction in New York’s history cannot agree on how many buildings changed hands.

The relevant number for the underwriting was neither of those. It was the rent-stabilised share, and that is disputed too: the Times said nearly three quarters at the time of sale, Multifamily Executive said about eighty per cent.^3,10 Either way, the overwhelming majority of the income was set by law rather than by the market, at rents far below what the same square footage would fetch a few blocks away. The value the buyers were paying for was not the rent roll. It was the difference between the rent roll and the market — and the legal mechanism that let an owner close that gap one apartment at a time.

That mechanism was vacancy decontrol, in force since 1993. When a rent-stabilised apartment came empty and its legal rent crossed a statutory threshold, it left regulation permanently. Renovation spending and vacancy bonuses were the tools for pushing a legal rent over the line. So the plan was to empty apartments, improve them, deregulate them and re-let them at market — and the seller’s own offering memorandum said how fast. Prepared by MetLife and CBRE, it targeted a rent-stabilised share below thirty per cent by 2018 and projected net operating income reaching $504.1 million by 2017.¹²

MetLife had already been running that play, gently. It told the Times at the time of sale that rents on twenty-seven per cent of the apartments were at market, and that another 1,800 units would be decontrolled over the following two years.³ What it was selling, in other words, was a conversion already in progress and a projection of how much faster somebody else could run it.

Net operating income in 2006 was $112 million.⁹ The memo was projecting a four-and-a-half-fold increase inside eleven years, and the buyers paid a price consistent with believing a version of it. The sign on the tower at 14th Street and First Avenue already said luxury rentals; the rent rolls said something else entirely.

Ten weeks before the sale closed

The sign already said luxury rentals. The rent rolls said something else entirely.

Stuyvesant Town and Peter Cooper Village from 1st Avenue and 14th Street on 3 September 2006, with a leasing sign for both properties painted on the building face — ten weeks before the sale to Tishman Speyer closed.Photo: David Shankbone, CC BY 2.5, via Wikimedia Commons

Forty per cent, from day one

Here is the correction that matters most, because the wrong version of it circulates everywhere. The figure usually quoted alongside the $112 million of income is $165 million of annual senior debt service. That number appears in no source this brief could find. It is reproducible as arithmetic — a plausible rate on $3 billion — but arithmetic that reproduces a figure is not a source for it, and a number nobody published does not become one by being repeated.

What is sourced, twice and independently, is the coverage ratio. Gothamist, reviewing the deal in 2016, wrote that rental income covered a scant forty per cent of the new, inflated debt load.⁹ City Journal, reviewing Charles Bagli’s book on the transaction, put it identically: from the first day, income from the property covered only forty per cent of the debt service.⁸ Two writers, two sources, one number — and it is a more damning number than the invented one, because it is a ratio rather than a dollar figure and therefore cannot be argued down.

A property covering forty per cent of its debt service is not a property with a financing problem. It is a property whose financing assumes a future that has not happened yet. The difference had to come from somewhere for as long as it took the plan to work, and the somewhere was a reserve fund carved out of the closing.

The lawsuit was filed three months after closing

In January 2007, nine residents of seven apartments sued. Their argument turned on J-51, a city tax abatement for building improvements — asbestos removal, boiler replacement, gut rehabilitation — which in exchange required the apartments to be registered with the state and kept under rent stabilisation for as long as the benefits ran.¹³

The 1993 statute that created luxury decontrol carved out an exception: it did not apply to apartments that became subject to rent stabilisation “by virtue of” receiving J-51 benefits. The state housing agency and every landlord had read that for years to meansolely by virtue of — so a building already stabilised for other reasons could take J-51 money and still deregulate. The tenants read it literally. The entire case, and with it the business plan, was four words in a statute.¹³

Justice Richard B. Lowe III dismissed the complaint in August 2007, adopting the agency’s reading.¹³ The Appellate Division reversed unanimously in March 2009.¹⁴ And on 22 October 2009 the Court of Appeals affirmed the reversal, four to two, in a per curiam opinion that dealt with the consequences in a sentence: defendants predict dire financial consequences from our ruling, and if the statute imposes unacceptable burdens, their remedy is to seek legislative relief.¹³

The court expressly left open whether the ruling applied retroactively, noting in a footnote that the lower courts had not considered the question.¹³ That single unresolved issue is what made the decision unpriceable rather than merely expensive — nobody could say what the liability was. Cleary Gottlieb told clients within a week that a property whose income had been projected to triple by 2011 and whose value had been projected at $7 billion within five years now looked highly unlikely to reach either.¹⁶ The Times reported the same day that the ruling could affect landlords of as many as 80,000 apartments across the city.¹⁵

It took another three and a half years to settle. In November 2012 the parties signed: $68.75 million of cash for overcharges running from January 2003 through December 2011, paid $58.25 million by the ownership entities and $10.5 million by MetLife’s successor company.¹⁷ Final approval came on 10 April 2013 — from Justice Lowe, the same judge who had dismissed the case six years earlier. The settlement returned 4,311 apartments to rent stabilisation through June 2020 and put total tenant recovery at $173.25 million, with legal fees of $18.9 million, under eleven per cent.¹⁸ The two press releases from the same firms six months apart give the class as 21,250 members and then about 27,500, and never reconcile the difference.^17,18

The reserve ran out first

The reserve is where the two most confident accounts of this deal disagree with each other, and neither is a bad source. The New York Times, reporting contemporaneously in October 2009 and again in January 2010, describes a single fund originally stuffed with $890 million for capital improvements, interest payments and renovations.^6,15 Multifamily Executive, writing in September 2009 with more granularity, describes two: a $400 million debt-service reserve set at issuance, down to $56.5 million by July 2009 and expected by Moody’s and Fitch to be exhausted by year end, plus a separate $190 million general reserve for capital improvements and debt paydown that had already run out somewhere between October 2008 and September 2009.¹⁰ Four hundred plus one hundred and ninety is $590 million, which is also the figure Commercial Property Executive uses.¹¹

This brief does not resolve that, because no source does. What both accounts agree on is the shape: a fund built to cover the gap between income and debt service, drawn down every month the plan failed to deliver, with no mechanism to refill it. Realpoint’s Frank Innuarto told the Times in October 2009 that about $24 million was left.¹⁵ Trepp’s Manus Clancy put the underwriting problem in one sentence in the same article: the property has never been able to generate nearly enough cash to service its debt, the interest reserve is dwindling, and now the property faces the prospect of taking a hit on its income.¹⁵

The loan went to CWCapital as special servicer in November 2009.¹⁹ On 8 January 2010, Tishman Speyer and BlackRock confirmed in a joint statement that they would miss that day’s scheduled $16 million payment.^20,21 Three years and seven weeks after closing.

The handback was not a moment

Almost every retelling of this deal has a scene in it where the owners hand over the keys in January 2010. The scene is real as journalism and wrong as a legal event, and the difference is worth four and a half years.

What happened in January 2010 is that the partnership stopped paying and said publicly it would give the property up. NPR and Marketplace, both on 25 January 2010, describe an agreement to turn the properties over to the lenders.^22,23 No legal instrument is named in any contemporaneous source, because none was executed. Title did not move. The borrower entities kept holding the asset while CWCapital, as special servicer for the senior bondholders, ran it.

The only formally documented conveyance in the entire collapse is dated 3 June 2014, when CWCapital executed a deed in lieu of foreclosure, took direct title, and cancelled a UCC foreclosure auction that had been scheduled for ten days later.³⁵ Transfer taxes on that conveyance ran to more than $100 million by one account and over $130 million by another.³⁵ And there was never a receivership: no court-ordered receiver appears anywhere in the 2010–2014 record. A property widely described as having been surrendered to its lenders in 2010 was legally still owned by the defaulted borrower until the summer of 2014.

Four and a half years of fighting over the wreckage

The gap between the informal surrender and the actual conveyance is where the capital stack did its real damage, because eleven levels of mezzanine debt is eleven levels of people with standing to sue.

The most instructive fight was Bill Ackman’s. In August 2010, Pershing Square and Winthrop Realty Trust, through a joint venture called PSW NYC, bought the first three mezzanine loans — $300 million of face value — for $45 million, about fifteen cents on the dollar.^24,37 The plan was elegant: mezzanine debt is secured by the equity in the holding company rather than by the real estate, so foreclosing on it would hand PSW control of the entity above the senior lenders’ heads, after which the complex could be converted to a co-op.

On 16 September 2010 Justice Richard Lowe — again — enjoined the foreclosure, holding that under the intercreditor agreement PSW would first have to pay off the $3.67 billion owed to the senior lenders.^26,37 The appeals court denied a stay at the end of the month.²⁷ On 26 October CWCapital settled by paying PSW $45 million and taking the loans back.²⁸ Pershing Square recovered exactly what it had paid. The cleverest trade in the whole story returned precisely zero, because an intercreditor agreement signed in 2006 said it would.

A correction is owed here to a claim that circulates widely: Appaloosa Management held senior CMBS bonds, not mezzanine debt. It was on the top of the stack, not the bottom. Appaloosa sued unsuccessfully in 2010 to force a bankruptcy rather than a foreclosure, arguing that route would save lenders roughly $200 million in transfer taxes, and sued again in November 2015 over CWCapital’s claim to as much as $566 million of default interest out of the sale proceeds — bondholders said the right figure was $38.2 million.^39,40,49 That second case was withdrawn on 1 December 2015 after a judge denied an injunction on standing, so it was never decided on the merits.⁴¹

The servicer’s own position deserves the stain rule. CWCapital was controlled by Fortress Investment Group. It installed a manager, then replaced that manager with CompassRock, a sister company under the same parent, effective 1 September 2012, citing a 33 per cent cut in management fees.^11,32 It rebuffed a tenant-ownership proposal backed by Brookfield and cancelled a scheduled foreclosure auction without explanation, and residents accused it of stalling a sale to protect its fee stream.³³ In May 2014, while CWCapital was moving to foreclose on the junior debt, its parent Fortress was separately reported to be assembling a $4.7 billion bid of its own — which drew a public conflict-of-interest objection from Councilman Garodnick.³⁴ CW’s counsel later described a $7.5 million annual special servicing fee across nearly six years and a $15 million liquidation fee on the sale.⁴⁹

In July 2014 six Centerbridge-affiliated junior lender entities sued, alleging the deed in lieu rested on a false premise: that CWCapital had represented $4.4 billion as owed on the senior mortgage when the true figure was around $3.45 billion, and that a property believed to be worth approximately $5 billion had been taken in a way that froze the juniors out of more than a billion dollars of recovery. CW’s spokesman called it purchased litigation.³⁶ It settled in early September 2015 — which is what cleared the path for a sale six weeks later.

What the equity lost, and the one total anybody published

The losses are documented investor by investor and never in aggregate by anyone with a ledger. CalPERS put in $500 million, wrote it off, and then fired BlackRock from the roughly $1 billion apartment portfolio it managed on the fund’s behalf.²⁹ The Florida State Board of Administration put in $250 million in 2007; its director Ash Williams told trustees on 1 September 2009 that the fund was carrying the investment at zero because the market had softened dramatically — an anticipatory markdown four months before the default.³¹ The Church Commissioners for England lost £40 million, less than one per cent of a then-£4.4 billion portfolio, and described the mechanism precisely: the value of the property fell to the extent that it covered only part of the borrowings, and the rest of the borrowings and the investor’s equity were lost.³⁰

Exactly one aggregate figure exists in print. Anthony La Malfa of BDO, quoted in Real Estate Weekly on 16 December 2015, put it at $2.4 billion in losses for the original equity and mezzanine investors.⁴¹ That is one analyst’s arithmetic in one trade publication, and it is the only total anyone has published. It should not be repeated without his name attached to it, and this brief does not repeat it without one.

The senior lenders did fine. The $3 billion senior CMBS loan was reported paid off after the December 2015 closing, with analysts estimating a gain on the proceeds above principal in excess of $1.1 billion.⁵⁰ Being at the top of an over-levered stack is a different business from being anywhere else in it.

The second sale, and what the city bought

CWCapital hired Eastdil Secured and ran a sale. The contract was signed on 19 October 2015, announced with Mayor de Blasio the following day, and closed on 18 December.^42,44,48 The price has three versions in circulation and one source resolves them: $5.3 billion is the net negotiated price, and the $5.45 billion figure reported at closing includes full payment of transfer taxes. Real Estate Weekly is the only outlet that explains the gap, and it did so by asking Blackstone, whose spokesperson confirmed the net price was still $5.3 billion.⁴⁹ Figures of $5.4 billion, $5.5 billion and $5.6 billion all appear in otherwise careful reporting, mostly through conflation with the 2006 price.

The financing was announced on 7 December: $2.7 billion, ten years, from Fannie Mae through Wells Fargo Multifamily Capital, and Fannie called it low-leverage in its own announcement.⁴⁷ Against a $5.3 billion price that is an implied loan-to-value near fifty-one per cent, though no source states one. Set beside the 2006 stack — 55.6 per cent of senior alone, with $1.4 billion of mezzanine on top of it — the difference is not the senior loan. It is everything above the senior loan.

What made the sale politically possible was a twenty-year agreement announced alongside it. Five thousand of the roughly 11,241 apartments were locked below market for twenty years, with about 1,400 “Roberts” units getting five additional years of protection and annual increases capped at five per cent from 2020, when the J-51 abatement expired.⁴⁴ The city’s own release set the income bands: ninety per cent of the covered units for households earning no more than $128,210 for a family of three, ten per cent for households at no more than $62,150, and no new tenant paying more than thirty per cent of income in rent.⁴⁴ Translated into rents by the reporting at the time, that is about $3,205 a month for a two-bedroom at the upper band and up to $1,553 at the lower.⁴⁵ Blackstone also agreed not to build new units on the site and not to convert to condominiums, and covered stabilised apartments were to be re-rented to income-qualified tenants at restricted rents on vacancy rather than deregulated.⁴⁴

The city’s contribution was reported as $225 million, $221 million and $220 million by different outlets and has two components. The larger is a $143,718,750 subordinate acquisition loan from the New York City Housing Development Corporation: twenty years, zero per cent interest, and forgiven annually at $7,185,937.50 — which is to say never repaid. The Real Deal reviewed the term sheet directly, and a mayoral spokesperson explained the structure as an enforcement device, on the grounds that structuring it as a loan gives the city significantly more authority if the affordability agreement is broken.⁴⁶ The smaller is a $77 million mortgage recording tax waiver.⁹

A third element got less attention. The term sheet included the city’s agreement to support transferring more than a million square feet of unused air rights off the site, in exchange for the no-new-construction and no-condo commitments — a clause Gothamist noted was in the documents but went curiously unmentioned at the press conference.⁹

And the stain on the public side is the city’s own auditor. The administration credited the deal with 100,000 apartment-years of affordability. In January 2018 the Independent Budget Office found that most of the units in question would have remained rent-regulated anyway, and credited the agreement with 36,000 apartment-years of additional affordability — roughly a third of the claim.⁵¹ Deputy Mayor Alicia Glen had told the 2015 announcement that the deal was saving taxpayers a lot of money. The IBO report as published could not be retrieved as readable text for this brief; the finding is taken from Patch’s contemporaneous account, which quotes it directly.⁵¹

Same eighty acres, both times

Nothing about the buildings changed. Everything about the capital did.

A tree-lined path inside Stuyvesant Town in August 2008 — between the two sales, while the first one was failing.Photo: David Shankbone, CC BY-SA 3.0, via Wikimedia Commons

Then the law removed the plan entirely

Blackstone bought with a residual version of the same upside every previous owner had underwritten: the roughly six thousand apartments not covered by the affordability agreement could still be deregulated as they came empty, and the Roberts protections expired in June 2020. In June 2019 the state legislature passed the Housing Stability and Tenant Protection Act and closed that door.

The Second Circuit’s own summary of the statute is the clearest one available. HSTPA repealed vacancy decontrol and high-income decontrol outright — both 1993 exits, gone. It repealed the statutory vacancy bonus and the longevity bonus. It capped recoupment of individual apartment improvements at $15,000 in aggregate over fifteen years with no more than three increases in that window, and made those increases expire after thirty years rather than becoming permanent. It capped major capital improvement increases at two per cent, amortised them over twelve to twelve and a half years, and restricted eligibility.⁵³ The exits were removed and the tools for reaching them were shrunk at the same time.

Here the third correction is due, because the popular telling collapses two entirely separate lawsuits into one. The federal case is a constitutional takings challenge to HSTPA filed in July 2019 by the Community Housing Improvement Program and the Rent Stabilization Association, with a companion as-applied case. It was dismissed at first instance, affirmed by the Second Circuit on 6 February 2023 under the Penn Central test, and the Supreme Court denied certiorari on 2 October 2023 and again on 20 February 2024.^53,54,55 Blackstone was never a party to it. It does not appear in the Second Circuit’s caption, and no source shows it funded the challenge.

The Stuy Town case is a different one, in state court. The tenants association sued Blackstone arguing that under HSTPA and the property’s J-51 history, about 6,200 apartments the owner planned to deregulate from mid-2020 had to stay stabilised. Justice Robert Reed ruled for the tenants, reported in early January 2023.⁵² Blackstone appealed, and then withdrew the appeal by filed stipulation on 23 and 24 February 2024 — days after the second cert denial, and unrelated to it.^56,57 So the accurate sentence is that Blackstone dropped its own appeal of a state-court ruling in the tenants association’s lawsuit in late February 2024, not that it abandoned a Supreme Court case it was never in.

The consequence is total. Every apartment in Stuyvesant Town and Peter Cooper Village is now rent-stabilised.⁵⁹ The plan that both the 2006 buyers and the 2015 buyers underwrote — convert roughly half the units to market rents — no longer exists as a legal possibility. Blackstone’s stated cumulative capital investment in the property has been reported four different ways from the company’s own statements and its reporting: $300 million in January 2023, $375 million in February 2024, more than $425 million and $460 million in December 2025 pieces published a day apart.^57,58,59

What happened when the loan came due

The reel this brief grew from was posted on 14 December 2025 and framed the refinancing as a test still to come. It had already happened. The $2.7 billion Fannie Mae loan matured in December 2025, and a $3.15 billion mortgage originated by Wells Fargo closed on 26 November 2025 and was recorded the following week.^58,59 The new loan is larger than the one it replaced, which is what a lender does when it is comfortable, in a property class where comfort has been scarce.

Scarce is the word. Using Ariel Property Advisors’ transaction data, buildings that are at least seventy-five per cent rent-stabilised traded at a citywide average of $270,000 a unit in 2019 and $163,000 a unit in 2025 — a forty per cent decline. Price per square foot fell from $399 to $209, and Manhattan’s per-foot decline was sixty-one per cent.⁶⁰ Cumulative operating expenses for stabilised buildings rose about forty per cent since 2019 while the rent increases the Rent Guidelines Board approved totalled sixteen per cent — a twenty-four point gap that compounds every year it persists.⁶¹

Against that, no current appraised value, occupancy figure or net operating income for Stuy Town has been published for 2024, 2025 or 2026, and no write-down attributable to HSTPA appears anywhere in the record. There is no evidence the property is being marketed. What can be said with sources is narrower and still meaningful: a lender wrote a bigger cheque against the same asset ten years on, in a sector whose average valuation fell by forty per cent over the same period.

What transfers

Sponsor equity is not a fee, it is a brake. Fifty-six million dollars against a six-billion dollar bill meant the people making the decisions had almost nothing to lose by being wrong, and the Times said so plainly in October 2009: Tishman Speyer itself stood to lose only the $56 million it had put in, while its partners, investors and lenders would take a far bigger hit.¹⁵ When you are looking at a capital stack, the useful question is not what the leverage ratio is. It is what the sponsor loses in the bad case, and whether that is enough to make them refuse the deal at the price.

A business plan that depends on a legal interpretation is a legal position, not a real estate position. The whole $5.4 billion rested on reading “by virtue of” to mean “solely by virtue of”, which the state housing agency and the entire industry had done for a decade and a half. Underwriting an entitlement — a rezoning, an abatement, a deregulation pathway — means underwriting an outcome that a court or a legislature can reverse without compensating anybody, and the reversal arrives on its own schedule rather than the fund’s.

Interest-only debt does not defer risk, it concentrates it at maturity, and a reserve fund is a countdown clock with a number on it. The 2006 structure had no amortisation and no mechanism to refill the reserve; the moment the conversion pace fell behind plan, the only question left was the date. Anyone can compute that date from the balance and the burn, and it should be the first thing computed rather than the thing discovered in a servicer’s report.

And complexity in the stack is a cost that only shows up in the bad case. Eleven levels of mezzanine debt bought nothing extra when things were fine and produced four and a half years of injunctions, competing foreclosures, allegations of a nine-hundred-million-dollar misstatement and a litigation settlement before the asset could be sold at all. A simpler stack cannot save a bad basis. A complicated one guarantees that a bad basis takes years to resolve, and that everyone junior discovers what an intercreditor agreement actually says at the worst possible moment.