What were the structural failures behind the 2008 subprime mortgage collapse and the financial crisis that followed? Separate the immediate trigger from the structural cause, and explain the mechanism rather than only the sequence of events.

The proximate cause was underwriting, but the scale came from the pipeline. Subprime rose from roughly 8% of U.S. mortgage originations in 2003 to about 20% in 2005 and 2006 — on the order of $600 billion a year — and roughly three-quarters of it was securitized, much of it as 2/28 adjustable-rate loans whose payment resets were printed in the prospectus. Global CDO issuance went from about $157 billion in 2004 to more than $500 billion in 2006. The alchemy was the re-slicing: a mezzanine CDO could buy BBB-rated tranches of subprime RMBS and reissue 70-80% of the resulting structure as AAA, which held only if defaults across regions stayed near-uncorrelated. That correlation assumption, embedded in the Gaussian-copula conventions the desks priced with, was the entire trade — and the agencies blessing it were paid by the issuers, with structured finance reaching roughly 44% of Moody's revenue by 2006. The Financial Crisis Inquiry Commission later found that Moody's downgraded 73% of the mortgage securities it rated AAA in 2006 all the way to junk. Fragility came from funding and leverage more than from the loans. The SEC's April 28, 2004 decision to let the largest broker-dealers use internal models for net capital helped push leverage past 30:1 — Lehman sat near 30.7:1 at the end of 2007, financing roughly $200 billion through tri-party repo that rolled overnight. AIG Financial Products, run by Joseph Cassano out of London, had written around $527 billion notional of credit default swaps, including some $78 billion of super-senior protection on multi-sector CDOs, against essentially no reserves, on the reasoning that AAA meant collateral would never be called. Meanwhile a public daily price on precisely this risk existed from January 2006 in Markit's ABX.HE indices: the BBB- 2006-1 series traded near par through 2006, broke into the low 60s by late February 2007 after HSBC's February 7 warning of a $10.5 billion charge and New Century's restatement that same week, and reached single digits by late 2007, with AAA ABX following in 2008. The ratings and the traded price disagreed for more than a year, and every capital rule in the system deferred to the rating. The sequence itself was legible in advance. Case-Shiller's 20-city composite peaked at 206.52 in July 2006 — 26 months before Lehman — and eventually bottomed near 134 in March 2012, roughly 35% down, with the national index off about 27%. New Century filed Chapter 11 on April 2, 2007; two Bear Stearns credit funds collapsed that June; Bernanke had told Congress on March 28, 2007 that subprime problems seemed "likely to be contained"; Chuck Prince gave the Financial Times his "as long as the music is playing" line on July 9, 2007. Moody's downgraded 399 subprime RMBS and S&P put 612 classes on watch on July 10, 2007. BNP Paribas froze three funds on August 9, 2007 citing a "complete evaporation of liquidity," and the ECB injected €94.8 billion that day. S&P then took action on roughly $534 billion of RMBS and CDO paper on January 30, 2008 alone. Bear Stearns went to JPMorgan in March 2008 at $2, later raised to $10, behind a $29 billion Fed backstop; Fannie and Freddie entered conservatorship on September 7, 2008; Lehman filed on September 15 with $639 billion in assets; AIG took $85 billion on September 16, eventually about $182 billion; the Reserve Primary Fund broke the buck that same day on $785 million of Lehman paper, triggering hundreds of billions in prime money-fund redemptions. The House voted down TARP 228-205 on September 29 (the Dow's 777.68-point drop was then its largest ever), the $700 billion bill was signed October 3, $125 billion went into nine banks on October 13, and the TED spread peaked at 4.58% on October 10. The S&P 500 fell from 1,565 on October 9, 2007 to 676.53 on March 9, 2009, about 57%, while unemployment went from 4.4% to 10.0% in October 2009 and the FCIC put household wealth losses near $11 trillion. The sharper lesson concerns what the AAA label actually was. Capital requirements, repo haircuts, investment mandates and CDS collateral triggers were all keyed to the rating, so a downgrade was not news about collateral — it was a synchronized, mechanical margin call on every holder at once. The bet underneath was single-factor and national, so the diversification was notional from the start; leverage supplied the transmission; the rating supplied the clock. Any system that lets one private label set collateral value across thousands of separate balance sheets has engineered a correlated liquidation into its plumbing.

This is the strongest of the four posts because it correctly identifies interconnection and leverage, not just bad loans, as the actual systemic mechanism — AIG wasn't holding mortgages, it was holding the counterparty risk of everyone who thought they'd hedged. The 30:1+ leverage figure is right but could use a companion stat: the shadow-banking repo market's overnight reliance on collateral that got repriced daily is what turned a housing correction into a liquidity crisis within weeks. One underweighted angle: the post treats this as a one-time event, but the same rating-agency-conflict-of-interest structure (issuer pays for the rating) hasn't fundamentally changed since 2008, which is worth noting for anyone using "it has an AAA rating" as a substitute for their own diligence today.

The information wasn't hidden, it was declined — Raghuram Rajan laid out the tail-risk case at Jackson Hole in 2005 and was largely brushed off, Fed governor Edward Gramlich spent years warning about subprime underwriting while the Fed sat on HOEPA rulemaking authority it had held since 1994 and didn't exercise until July 2008, and Brooksley Born's attempt to bring OTC derivatives under CFTC oversight was killed and then locked shut by the Commodity Futures Modernization Act of 2000. That's why the winners were the handful of people who read loan-level tapes instead of ratings: Michael Burry at Scion, who began buying subprime CDS in 2005, Steve Eisman at FrontPoint, Cornwall Capital's Charlie Ledley and Jamie Mai, Greg Lippmann marketing the short from inside Deutsche Bank, and John Paulson, whose funds cleared roughly $15 billion in 2007. A second class of winner gets counted less often — the balance sheets that still had cash when nobody else did: Buffett's $5 billion into Goldman preferred at a 10% dividend, JPMorgan taking Bear Stearns, Barclays buying Lehman's North American broker-dealer out of bankruptcy for a fraction of book, Wells Fargo absorbing Wachovia. The prevention that actually stuck is structural rather than predictive: Dodd-Frank's Title VII forced standardized swaps into central clearing with posted margin, so a $527 billion CDS book like AIG's can no longer be assembled bilaterally and uncollateralized, and Basel III set common equity tier 1 at 4.5% plus a 2.5% conservation buffer, added a leverage ratio that ignores risk weights entirely, and layered on the liquidity coverage ratio. None of that predicts the next bubble — it just means 30:1 leverage against AAA paper you didn't diligence is no longer a legal business model at a systemically important bank.