The Big Short

Michael Lewis · 2010

Thesis

The financial crisis of 2008 was built into the system before it happened: Wall Street turned bad home loans into securities rated as safe, the rating agencies and regulators failed to see it, and the few people who understood what was happening, mostly outsiders and misfits, saw it because they looked at what the securities actually contained.

The main ideas

Subprime mortgages were packaged into bonds, and the riskiest slices of those bonds were repackaged into new securities rated as safe.
The rating agencies' models could be gamed, and the agencies were paid by the banks whose products they rated.
A handful of investors bet against the market by buying credit default swaps, insurance against the bonds failing.
They saw the crisis coming because they read the details that others ignored.
Being right early was nearly as painful as being wrong: the bets cost money for years, and investors rebelled.
The incentives of everyone involved, from mortgage brokers to bond traders, rewarded making the loans and selling the securities, not their quality.

The machine

Michael Lewis, whose Liar's Poker (1989) described Wall Street in the 1980s, opens by recalling that he wrote that book as a warning, only to find that students read it as a manual. The Big Short (2010) describes the machine that, two decades later, produced a far larger disaster.

Its raw material was subprime mortgages, loans to borrowers with weak credit, often with low introductory rates that later reset higher. Thousands of these loans were pooled into mortgage bonds and divided into tranches, slices ranked by who would absorb losses first. The lowest slices, rated triple-B, were risky. Wall Street then pooled those risky slices into collateralized debt obligations, CDOs, and the rating agencies rated most of each CDO triple-A, on the theory that the underlying loans were diversified. In fact they were all bets on the same housing market.

Synthetic CDOs, built entirely from credit default swaps rather than actual loans, multiplied the bets beyond the mortgages that existed.

Implication: when a security is assembled from parts that are individually risky, ask what happens to all the parts at once in a downturn, not what each does on average.

Gaming the ratings

Lewis shows how the rating agencies, Moody's and Standard & Poor's, were outmaneuvered. Their models looked at the average credit score of the borrowers in a pool, not the distribution, so lenders could include very low scores balanced by higher ones. Borrowers with thin credit histories, such as recent immigrants, could have deceptively high scores.

The agencies were paid by the banks that issued the bonds, and the banks' traders understood the models better than the analysts who ran them. The bankers learned exactly how much risk they could put into a pool while keeping its rating.

The mechanism is that a rating system paid by those it rates, using models those it rates can study, invites optimization against the model rather than honest assessment.

Implication: treat a rating or certification as evidence about the rater's model, and ask who pays for it and whether the rated party can game it.

The outsiders

Lewis tells the story through the few who bet against the market. Michael Burry, a doctor turned investor who ran a fund called Scion Capital, read the prospectuses of individual mortgage bonds and concluded that the loans were bound to fail. In May 2005 he began buying credit default swaps on subprime mortgage bonds, insurance that would pay if the bonds defaulted.

Steve Eisman, who ran a fund at FrontPoint Partners, had long distrusted the subprime industry and, with his colleagues Vincent Daniel and Danny Moses, bet heavily against it. Greg Lippmann, a trader at Deutsche Bank, made the case for the bet to hedge funds and arranged the trades. Charlie Ledley and Jamie Mai, who had started Cornwall Capital with $110,000, bought cheap insurance against the higher-rated tranches of CDOs, which the market priced as nearly risk-free.

The mechanism common to all of them is that they examined what was inside the securities. The people they bet against relied on ratings and recent history.

Implication: in any market, the edge often lies in reading the underlying details that others accept on trust.

The cost of being early

Being right was not enough. Burry's swaps required premium payments year after year while housing prices kept rising. His investors, seeing the fund lose money, tried to withdraw, and he restricted withdrawals by placing the swaps in a separate account, provoking fury. Lewis shows the other short sellers enduring similar doubt.

When the crisis came, the bets paid enormously. Scion returned about 489 percent from its founding in 2000 to mid-2008, against just over 2 percent for the S&P 500; Cornwall turned its small start into tens of millions.

The mechanism is that a bet against a bubble pays only when the bubble breaks, and until then looks like a mistake to everyone, including the bettor's own investors.

Implication: if you take a contrarian position, make sure you can survive, financially and in the eyes of your backers, for longer than you expect to be wrong.

Who lost

On the other side of the trades were some of the largest firms on Wall Street. Howie Hubler, a trader at Morgan Stanley, had bet against the riskier mortgage tranches but financed the bet by selling insurance on supposedly safe triple-A CDO tranches; when those collapsed too, Morgan Stanley lost about $9 billion. AIG's Financial Products unit had insured large amounts of these securities; Lewis recounts how one of its executives, Gene Park, realized the bundles were mostly subprime, and how the unit stopped writing new insurance in 2006 but kept what it had already sold.

The book ends with Lewis taking John Gutfreund, his former boss at Salomon Brothers, to lunch. Gutfreund tells him that Liar's Poker destroyed his career and made Lewis's.

Implication: the largest losses fell on firms that believed their own models and were paid for volume rather than for judgment.

Limits

It tells the story through the winners. Its heroes are the few who profited; the book has less to say about the many who lost, and one of the people it portrays, the CDO manager Wing Chau, sued Lewis for defamation, though the courts ruled the statements were opinion.
Its account of causes is contested. Critics argued that it underplays government policy and the role of Fannie Mae and Freddie Mac; Lewis replied that the crisis was mainly the free market going wrong, and some studies found private-label securitization mattered more than the government-sponsored firms.
It concentrates on the mortgage market. The wider causes of the crisis, such as global capital flows and bank funding structures, are outside its scope.

When Genius Failed — an earlier collapse caused by leverage and trust in models.
The Black Swan + Antifragile + Skin in the Game + The Bed of Procrustes — why risk models built on recent history fail in rare crises.
Dot.con — the previous bubble, and the conflicts of interest that drove it.