Red Lobster was bought for $2.1B — and $1.5B of it came from selling its own restaurants

Ten years later the company filed for bankruptcy owing $190.5 million a year in rent, on leases its own court filing describes as “priced above market rates”.¹⁵ That is the story as it is usually told, and the arithmetic holds up: every number the reel version carries survived checking, including the twenty-five-year terms, the two per cent compounded escalators and the bundled leases.

The part that does not hold up is the causation, and it fails in an interesting direction. The phrase “sale-leaseback” does not appear anywhere in the 124-page declaration Red Lobster’s chief executive swore to on the day it filed.¹⁵ The company names the rent as a problem and never names the deal that wrote it.

The activists asked for this

The idea did not come from the buyer. In October 2013 Barington Capital went public with a letter urging Darden to break itself up and separate its real estate, which it valued at up to $4.4 billion, and named a sale-leaseback as one of the routes worth exploring.⁵ Two months later Starboard Value disclosed a 5.6 per cent stake and called Darden’s own plan insufficient.⁶ By April 2014 Starboard had filed a detailed primer valuing the property at about $4 billion “and possibly far more”.⁷

Then Darden did a version of what they had asked for, and both of them attacked it. The objection was not to monetising real estate; it was that this particular structure moved the value to the buyer rather than to Darden’s own shareholders. Starboard’s arithmetic, published six days after the announcement, backed the $1.5 billion of property out of the $2.1 billion price and concluded that the restaurant business itself had gone for about $600 million — or, once the tax inefficiency of the structure was counted, that the value received was “a paltry $100 million, or less than 1x EBITDA”.⁸

Darden rejected that publicly, saying Starboard’s analysis contained $107 million of costs that were never incurred and undervalued the consideration by nearly half a billion dollars, and that the transaction was struck at about nine times trailing earnings against comparable deals nearer eight.⁹ In October, shareholders elected all twelve of Starboard’s nominees and removed the entire board.¹⁰ The argument was settled by vote rather than by evidence, and the numbers on both sides are still sitting in the record where each side left them.

The asset, and the liability

It was bought with its own buildings.

A Red Lobster exterior. The freestanding suburban restaurant on its own lot is the asset that was sold in 2014 and the rent that was owed on it afterwards.Photo: Paige Baird, CC BY 2.0, via Wikimedia Commons

What $2.1 billion was made of

A headline price is not a cheque. This one was assembled, and the assembly is the whole mechanism.

The lease was built not to be broken

The reel version of this story says the leases were bundled, so closing one restaurant is not as simple as ending one lease. That is exactly right, and the buyer said so itself, in the release announcing the closing. Its president, David S. Kay:“Our team structured the transaction with multiple homogeneous long-term master lease pools which are cross defaulted and contain restrictions on leverage and assignability.”²

Read that as a landlord protecting a cash flow and it is unremarkable. Read it as an operator’s constraint and it is the whole trap. Cross-default means the pool is one obligation: failing on the weak restaurants inside it puts the healthy ones at risk too, so the tenant cannot quietly shed the bad locations and keep the good. A twenty-five-year weighted-average term with two per cent compounded escalations means the rent only ever goes one way, on a clock nobody can stop.² Ten years on, the filing recorded 247 of the leased locations sitting inside master leases and 440 held individually — the master leases being the ones that cannot be picked apart.¹⁵

What the filing actually says

Then the story arrives at the part everyone knows, and the record turns out to be more careful than the retelling.

The shrimp are real. In May 2023 the former chief executive made Ultimate Endless Shrimp a permanent $20 menu item “despite significant pushback from other members of the Company’s management team”, and the company’s own filing puts the cost of that decision at $11 million — the same figure Thai Union disclosed in its own earnings.^13,15,17 The promotion was announced to the public that June as being “here to stay, all day, every day”.¹³

But $11 million does not bankrupt a company with $1.5 billion of annual sales, and the filing does not claim it did. What it claims is broader: “a difficult macroeconomic environment, a bloated and underperforming restaurant footprint, failed or ill-advised strategic initiatives, and increased competition”.¹⁵ Guest counts were down about thirty per cent against 2019. Cash had fallen from roughly $100 million in May 2023 to under $30 million months later.¹⁵

And under the heading Unfavorable Leases & Underperforming Locations, the filing says the thing this whole study is about, in one sentence:“A material portion of the Company’s leases are priced above market rates.”¹⁵

What it never does is join the two ends. Across 124 pages the words “sale-leaseback” do not appear once, and the 2014 transaction is mentioned exactly once, as a single sentence of corporate history: “In May 2014, Darden sold Red Lobster in a leveraged buyout transaction to Golden Gate Capital.”¹⁵ No causal language, no reference back to it from the leases section. The company told the court its rent was above market and declined to say where the rent came from.

That silence is worth sitting with rather than reading through. A first-day declaration is sworn, written by restructuring professionals, and read by every creditor in the case — including the landlords. It is not the document in which a debtor volunteers that its own former owner created the problem. The inference the press drew is reasonable and the filing will not make it, and an honest brief has to report both of those facts rather than pick the tidier one.

The record has a stain

On 13 May 2024, six days before the filing, 93 restaurants were closed and vacated — in the filing’s own words, deemed non-performing because of rent costs and/or financial performance.¹⁵ Staff at several locations reported learning of it the morning it happened, some of them having worked the Mother’s Day shift the day before. A liquidation auction of equipment from dozens of the closed restaurants went public in the same few days.

A WARN Act class action was filed in Florida on 17 May, two days before the petition; the Chapter 11 filing triggered an automatic stay, and the case was later dismissed without prejudice on procedural grounds rather than on its merits. Court filings put employment at roughly 36,000 people across the United States and Canada at the time of filing.¹⁵ By the time the company emerged in September it was about 545 restaurants and roughly 30,000 employees.¹⁹

The rent did not go away with the bankruptcy. In December 2025, more than a year after emergence, the company cut about a tenth of its corporate staff and around two hundred restaurant employees, and pointed at the legacy leases that survived the restructuring. The mechanism outlived the company that signed it and the process meant to clear it.

What transfers

A sale-leaseback is not a trick and it is not free money. It is an exchange of an asset for cash plus a fixed obligation, and the reason it is so often a bad trade for an operator is that the two sides of it are recognised at different times. The cash arrives at once, on the day of the deal, where everyone can see it. The obligation arrives a little at a time, for twenty-five years, in an operating line, growing at a rate written into the contract by someone whose job was to protect it.

The specific things that made this one dangerous are all readable in the buyer’s own press release, on the day it closed. Cross-defaulted pools remove the tenant’s ability to close its worst locations without endangering its best. A compounding escalator removes any relationship between the rent and the business’s ability to pay it. A twenty-five-year term outlasts the strategy, the management, the owner and — here — the first bankruptcy.

And there is a discipline in how to say what happened. The rent was real, it was above market by the company’s own account, and it plainly contributed. But the company under oath listed four other causes first and never named the transaction at all, and the sale-leaseback was demanded by activists before it was executed by a buyer. Judge a structure by what it does over its life, not by the headline on the day it is signed — and when the record declines to draw the line you want to draw, say that it declined.