On 15 September 2008, Lehman Brothers — a 158-year-old Wall Street investment bank — filed for
the largest bankruptcy in US history. The next morning, the Federal Reserve announced an
emergency
The answer runs through the same theme as
Through the mid-2000s, US mortgage lenders originated a growing volume of
subprime mortgages — loans to borrowers with weak credit histories, often at
low "teaser" rates that reset higher after a year or two. Rather than hold these loans, banks
packaged thousands of them into mortgage-backed securities (MBS) and then
re-sliced pools of those securities into
That AAA rating rested on one crucial statistical assumption: that mortgage defaults across the country were not very correlated — a wave of defaults in Florida shouldn't have much to do with a wave of defaults in Nevada, so a nationally-diversified pool should see only modest, smoothed-out default rates even in a downturn. Rating models were calibrated on decades of US housing data in which nationwide house prices had never fallen year-on-year. That fact was treated as reassuring history. It was actually just a small sample that had never yet seen the scenario about to occur.
A London-based unit called AIG Financial Products (AIGFP) ran a business
selling
Two decisions made this catastrophic rather than merely large. First, unlike a typical derivatives dealer, AIGFP mostly did not hedge this book — it simply sold protection and kept the premium, a directional bet rather than a market-making operation. Second, it held very little capital against the position, because the models said the risk was negligible. The whole business was, in effect, LTCM's lesson wearing an insurance-company costume: a strategy that is individually plausible becomes existential once it's leveraged against too thin a cushion and concentrated in one hidden common factor — here, the assumption that US house prices, nationwide, do not fall together.
Once US house prices actually fell nationwide starting around 2006–2007, the chain ran quickly and simply:
Notice that AIG did not need bonds to formally default to be in trouble. Standard CDS contracts require the protection seller to post collateral whenever the market value of the reference asset falls or the seller's own credit rating is downgraded — long before any default is confirmed. As CDO tranche prices fell and rating agencies downgraded both the CDOs and AIG itself in 2008, counterparties like major investment banks made collateral calls totaling tens of billions of dollars, far larger and far earlier than AIGFP's models had ever contemplated. AIG simply did not have that much cash on hand, and because it was counterparty to so much of the banking system's protection, its failure threatened to leave every one of those banks suddenly unhedged and facing huge losses at the same moment — the systemic risk that triggered the government's intervention.
A structural conflict of interest sat underneath the whole machine: the banks that issued CDOs paid the rating agencies to rate them — the "issuer-pays" model. An agency that rated deals too harshly risked losing that issuer's future business to a rival agency. This didn't require outright corruption to matter; a systematic, unconscious tilt toward generous ratings, repeated across thousands of deals, was enough to leave "AAA" stamped on securities that turned out to be far riskier than advertised. A handful of contrarian investors — most famously the hedge-fund manager Michael Burry, later the subject of the book and film The Big Short — read the underlying mortgage data directly, concluded the ratings were wrong, and bought CDS protection on subprime bonds themselves, profiting enormously when the ratings finally caught up with reality.
A common myth holds that the crisis proves a small corner of the market can never really threaten the whole system, on the theory that subprime mortgages must have been somehow enormous. They weren't: the entire US subprime mortgage market was on the order of a few hundred billion dollars — tiny next to global financial markets worth tens of trillions. The disaster wasn't the size of the underlying losses; it was leverage and interconnection turning a modest sector's losses into a global event. CDS contracts let AIG take on exposure many times its capital base, and that exposure ran through nearly every major bank as a counterparty. A loss that, held directly and without leverage, would have been a bad year for mortgage lenders instead became — amplified through derivatives and transmitted through counterparty relationships — the worst financial crisis since the Great Depression. The lesson isn't "avoid small risky markets"; it's "watch how leverage and interconnectedness can turn a small fire into a systemic one," the same amplifier that undid Barings and LTCM at smaller scale.