AIG Blamed a London Unit. The London Unit Was in Connecticut.

On September 16, 2008, the Federal Reserve agreed to lend up to $85 billion to a company that, eight months earlier, had carried the highest credit rating in the world.³ The story that hardened almost overnight was tidy and convenient: a swashbuckling derivatives desk in London had loaded the firm with toxic bets while the grown-ups at headquarters were looking the other way. It is a good story. It has a foreign accent, a villain, and a clean line between the gamblers and the innocents. Almost none of it survives contact with the record.

The official story is that a rogue London unit blindsided an innocent insurance giant. The real story is that AIG Financial Products sat in Connecticut, the costliest single loss came from a business AIGFP never touched, and the thing that detonated the crisis was the parent's own credit rating — the one asset AIG was supposed to be guarding most carefully.

The London unit was in Wilton, Connecticut

Start with the geography, because the geography is the whole sleight of hand. AIG Financial Products was incorporated in Delaware and headquartered at 50 Danbury Road in Wilton, Connecticut — a suburban office park forty miles from Manhattan.¹ It ran a busy floor in London, along with offices in Paris, Tokyo, Houston, and Hong Kong.¹ London was a branch, not the brain. Naming the crisis after London does something specific: it makes the danger sound foreign, distant, and detached from the people who actually owned it. A unit in Mayfair can blindside you. A subsidiary in a Connecticut office park, consolidated onto your balance sheet and trading under your name, cannot. It is you.

The 'rogue founder' part of the legend collapses just as fast. AIGFP was founded on January 27, 1987 by Howard Sosin and two former Drexel Burnham Lambert partners; the man later cast as the unit's reckless architect, Joseph Cassano, arrived merely as one of ten original staffers.² Sosin ran it until a falling-out in 1993. Tom Savage ran it after that. Cassano did not become CEO until 2001 — fourteen years in.² And for most of that history the unit was not a time bomb at all: from 1987 to 2004 it contributed over $5 billion to AIG's pre-tax income.⁸ Seventeen profitable years do not describe a rogue operation. They describe a prized one, which is exactly why the parent kept feeding it the one ingredient that turned it lethal.

What actually pulled the trigger

Here is the mechanism the London story can't explain. AIGFP wrote credit default swaps — guarantees that if certain bundles of mortgages went bad, AIG would pay. By mid-2008 the gross notional on those swaps was roughly $440 billion, concentrated in the super-senior slices of multi-sector CDOs.⁵ (The eye-popping $2.7 trillion figure that floats around is the unit's total derivatives book across every product line, not its CDS exposure — conflating the two inflates the danger and obscures where it actually lived.⁵) But notional is not the loss. The killer was a clause. Because AIG carried a top-tier credit rating, it had been able to write those swaps without posting collateral up front. The day the rating slipped, the contracts demanded billions in cash collateral — immediately. The very AAA that let AIGFP win the business was the fuse that, once lit, drained the whole company's liquidity at once. AIG had pledged its best asset as a tripwire and then walked across it.

And then there is the loss the legend forgets entirely. AIG ran a securities-lending program — lending out the high-quality bonds its insurance subsidiaries held, and reinvesting the cash collateral in mortgage-backed paper. When that paper soured and counterparties wanted their cash back, the program lost roughly $21 billion.⁶ That program had nothing to do with AIGFP. It was run inside AIG's insurance subsidiaries — the supposedly safe, boring core of the company.⁶ Set it beside AIGFP's CDS losses of about $30 billion and the picture changes shape completely: this was not one rogue desk. It was two separate businesses, in two separate parts of the firm, draining cash in the same direction at the same moment.⁶

The watchdog that watched the wrong thing

The other comforting half of the myth is that nobody could possibly have seen this coming — the exposures were too exotic, too hidden offshore. Not so. The Office of Thrift Supervision was AIGFP's designated consolidated supervisor and did conduct on-site reviews of the unit.⁷ The failure was not blindness; it was that a thrift regulator — an agency built to supervise savings-and-loans — had somehow ended up policing one of the largest derivatives books on earth, and it never grasped the liquidity risk it was looking straight at.⁷ When the OTS director admitted to the Financial Crisis Inquiry Commission that he 'had no clue' what AIGFP was doing as late as September 2008, he wasn't confessing that the information was unavailable.⁷ He was confessing that the firm had shopped its way into the gentlest available oversight, and the gentleness was the point.

Isn't 'AIGFP did it' close enough?

The fair objection is that AIGFP was, in fact, the largest single source of loss, and pinning the crisis on it is roughly right. There is truth there: the CDS book was the biggest hole, and Cassano's post-2001 push into mortgage-linked swaps was the strategy shift that turned a profitable unit into a lethal one.⁸ But 'roughly right' is exactly the trap, because where you place the blame determines what you fix. If the lesson is 'rein in the rogue desk,' you keep the AAA-as-collateral structure, the securities-lending machine inside your insurers, and the captured regulator — and you congratulate yourself on solving the problem. The honest reading is harder: the parent board approved the strategy, the parent's rating was the trigger, and a second loss of comparable size grew in the part of the company everyone called safe. The London frame is not a small inaccuracy. It is an alibi.

The numbers that survive are stark and, in the end, almost ironic. The initial Fed line was up to $85 billion; total government commitments climbed to roughly $182 billion across the restructurings that followed.³⁴ And the taxpayer, having rescued a company felled by its own best asset, eventually came out ahead — a positive return of about $22.7 billion on the whole rescue.⁴ AIG nearly died not because a foreign desk went rogue, but because it built its entire derivatives franchise on the assumption that its rating could never slip, ran a second mortgage bet in the room it called safe, and chose the regulator least able to notice. It out-sourced the blame to London. The risk had been at home the whole time.