AIG Didn't Have One Fire. It Had Two, Burning Toward Each Other.

On September 16, 2008, the Federal Reserve authorized a loan of up to $85 billion to a single insurance company at an interest rate of three-month LIBOR plus 850 basis points—a punitive rate, secured by every asset AIG and its non-regulated subsidiaries owned.¹ The day before, Lehman had been allowed to die. AIG was not. The difference was not size or sentiment. One explanation that gained currency among analysts was that AIG had quietly wired itself into the plumbing of the global financial system in two separate places at once, and both were on fire—though the precise reasoning behind the differential treatment has never been officially documented.

The story everyone tells is simple: AIG's London traders wrote a mountain of credit default swaps, the housing market fell, and the bets came due. It's a clean villain and a clean weapon. It is also only half the fire—and the half that got all the cameras.

Two books, both losing, neither watching the other

AIG Financial Products had written credit default swaps on over $500 billion in assets, including $78 billion on multi-sector collateralized debt obligations—insurance on bonds that, when the bonds soured, required AIGFP to post collateral it didn't have.³ That is the famous book. The unfamous one sat inside AIG's life insurance companies. To squeeze extra yield, they had lent out securities and reinvested the cash collateral into mortgage-backed assets. When markets froze, borrowers wanted their cash back and the reinvested assets couldn't be sold to fund it. Those life insurers took roughly $21 billion in securities-lending losses in 2008 alone.⁸

Here is the part that rewrites the legend. Federal Reserve Bank of Chicago and NBER research found that the securities-lending losses were of a similar magnitude to the CDS losses—cumulative losses from both activities ran on the order of $50 billion by the day of the rescue.³ AIG did not have a credit default swap problem with a securities-lending footnote. It had two structurally separate risk books, in two different corners of the company, draining liquidity in lockstep at the worst possible moment.

So this is the thesis, and it cuts against the textbook version. AIG's defining crisis was not a single catastrophic bet. It was a dual failure of two mutually reinforcing risk books that nobody at the top—or in regulation—was watching as one system. And the rescue that followed, however well it paid off, set a precedent that remains genuinely contested: it paid counterparties at or near par and was never fully held to account for it.

Insured by a company nobody had insured against

How does a Fortune 20 company carry two parallel time bombs without anyone pulling the alarm? The answer is a regulatory accident. AIG owned a small thrift subsidiary, and that one fact placed the entire holding company under the Office of Thrift Supervision—a regulator built to oversee savings-and-loans, not a global derivatives book. The OTS never grasped the liquidity risk inside AIGFP. A former OTS director told the Financial Crisis Inquiry Commission he had 'no clue' about the scale of the problem as late as September 2008.⁶ The watchdog assigned to the most dangerous insurance company on earth was looking at the wrong thing entirely.

The rescue that wasn't a bailout, and the bill that wasn't $182 billion

Almost everything the public 'knows' about the AIG bailout is slightly off, and the slippage matters. The first rescue was not TARP—it was a Federal Reserve loan under Section 13(3) of the Federal Reserve Act, the central bank's emergency-lending power, structured as a revolving credit facility secured by a pledge of all of AIG's material subsidiaries.¹² TARP's $40 billion in preferred stock arrived only with the November 2008 restructuring, which also cut the credit facility from $85 billion to $60 billion, stretched it to five years, slashed the rate to LIBOR plus 300 basis points, and spun up two special vehicles—Maiden Lane II and III—to absorb the toxic assets directly.⁵ The single number everyone repeats—$182 billion—was never the cost. It was the maximum combined commitment authorized: the ceiling.⁴

The haircut nobody negotiated

Here is where the clean win gets a stain. When AIGFP's swaps came due, its counterparties—Goldman Sachs, Société Générale, and others—were made nearly whole. The precise phrase '100 cents on the dollar' is technically wrong: the New York Fed later said AIG paid 'slightly less than 100 cents on the dollar' once you account for collateral already posted, which is exactly why that sentence was struck from a December 2008 SEC filing.⁷ But the substance survives the technicality. No meaningful haircut was negotiated. SIGTARP confirmed that Goldman simply would not agree to concessions, because a concession would have meant booking a loss.⁷ The Fed, holding the leverage of a counterparty's survival, declined to use it. Public money flowed through AIG and out to the very banks sitting on the other side of those mortgage-linked contracts.

But it made money — doesn't that settle it?

The honest objection is the strongest one: the rescue worked. It stopped a cascade, it stabilized the system, and it returned a documented $22.7 billion to the government.⁴ By the only metric that fits on a chart, this was among the most consequential crisis interventions of the era. Why pick at it? Because the headline profit quietly rests on more than market recovery. As former members of the Congressional Oversight Panel—including chair Elizabeth Warren—noted, a special tax exemption allowed AIG to carry forward net operating losses against future tax bills; some estimates put the resulting benefit to the company at $17.7 billion in profits, so the true net to taxpayers is materially lower than the round number suggests.⁹ The profit was real. It was also subsidized by a tax break the public never voted on, paid to counterparties who never took a haircut, ushered through a disclosure the Fed asked to be deleted. 'It made money' answers the accounting question and dodges the governance one.

AIG's defining crisis is remembered as a parable about derivatives, with a happy ending stapled to the back. It was something stranger and more instructive: a company that failed in two places at once, watched by a regulator built for a different business, rescued by a power most people had never heard of, and saved on terms that protected the people across the table from any pain. The taxpayer got the $22.7 billion. The precedent—pay par, strike the sentence, count the profit, move on—got everything else. The fire was put out. The question of who was allowed to start it, and who was made whole for it, is still smoldering.