John Paulson walked into the offices of Goldman Sachs and asked the bank to manufacture something the cash market could not supply on its own: bespoke exposure to the riskiest tranches of subprime mortgage bonds, packaged into synthetic securities he could then bet against at scale. He bought credit default swaps referencing those bonds across several dealer banks, insurance contracts that would pay off if the underlying mortgages defaulted, and he leaned on the banks to find the investors willing to take the other side. Within roughly a year, that bet paid his personal account close to four billion dollars, at the time the largest one-year payday any individual had ever pulled from financial markets.
Four years later, his funds had one of the worst years in their history. One of the losses, a stake in a Toronto-listed Chinese forestry company called Sino-Forest, became the symbol of the reversal: a position his team had bought with almost none of the verification work that had powered the mortgage trade, and one he sold within two weeks of a short-seller calling the company a fraud.
He still runs money from Manhattan. The fund is smaller now. The phone rings less.
The trade that printed four billion dollars
The mechanics of the subprime short were less exotic than the legend suggests. Paulson & Co. paid premiums on credit default swaps referencing BBB-rated tranches of mortgage-backed securities, insurance contracts that paid off if the underlying bonds defaulted. The annual cost was small. The payoff if the bonds went to zero was the full notional. Paulson built the position across 2006 and into early 2007, when the cost of that insurance was still cheap because the consensus view held that nationwide home prices had never fallen in modern memory and therefore could not fall now.
His firm ran the numbers differently. The team, led by Paolo Pellegrini, spent roughly eighteen months building what became the analytical core of the trade: a proprietary database of regional home prices going back decades, assembled from county-level deed records, OFHEO repeat-sales indices, and individual metro-area pricing series stitched together by hand. Pellegrini’s group ran the data in real terms rather than nominal, which produced a different picture than the one the rating agencies were working from. In real terms, home prices had detached from their long-run trend by a wide margin, and the regional record contained multiple sharp declines that the agency models had effectively averaged away.
From that base, the team built loan-level projections on the 2005 and 2006 vintages. They pulled remittance data on individual deals, tracked FICO distributions and loan-to-value ratios on the underlying collateral, modeled documentation quality by originator, and ran cash-flow waterfalls on specific BBB tranches under price-decline scenarios the agency models did not contemplate. The conclusion was that the BBB tranches would be wiped out entirely under price declines as modest as five to ten percent nationally, declines that, on the regional data Pellegrini had assembled, were not extreme at all.
When the ABX index tracking subprime mortgage exposure collapsed in 2007, Paulson’s flagship Credit Opportunities fund returned roughly 590 percent, and his funds collectively generated around fifteen billion dollars for investors. Paulson’s personal cut, after management fees and his own capital in the funds, came to nearly four billion dollars for 2007 alone, a figure Gregory Zuckerman later put at about $3.7 billion, among the largest single-year takes in Wall Street history. He has examined other unlikely contrarian bets in our coverage of market outliers and the people who make them.
The Abacus footnote
One of the deals Paulson’s firm helped shape became famous for the wrong reasons. The Abacus 2007-AC1 CDO was a synthetic security Goldman Sachs assembled with input from Paulson & Co. on the reference portfolio. Paulson then shorted the same instrument. According to the SEC, the bonds in the portfolio were downgraded almost in their entirety within months, investors in the CDO lost more than a billion dollars, and Paulson’s opposite position yielded a profit of roughly one billion dollars. The SEC charged Goldman, alleging the bank failed to disclose Paulson’s role in selecting the underlying bonds to the investors who bought the long side. Goldman settled for $550 million, then the largest penalty the agency had levied on a Wall Street firm, without admitting or denying the allegations, though it acknowledged its marketing materials had been incomplete. Paulson & Co. was never charged. In the SEC’s framing, the firm had owed no disclosure obligations to the long-side investors.
The episode is worth pausing on because it reveals something the popular narrative tends to flatten. Paulson did not simply read the housing market correctly. He commissioned the construction of the very instruments he then bet against, in volumes the natural market would not have supplied on its own. The asymmetry between his analytical work and the rating agency models was the edge. The willingness of the dealer banks to manufacture exposure to order was the mechanism.
The fund that became a juggernaut
By 2010, Paulson & Co. managed roughly 36 billion dollars, and it kept growing into 2011. The firm had become one of the largest hedge funds in the world. Paulson the person had moved from a profitable but mid-tier merger arbitrage manager into the same conversation as George Soros and Stanley Druckenmiller.
The next bet was gold. Paulson launched a dedicated gold fund in 2010 with share classes denominated in gold ounces, built largely on gold mining equities and bullion-linked derivatives, much of it his own money. The thesis was straightforward: the Federal Reserve’s response to the financial crisis would eventually produce significant inflation, and gold would be the cleanest hedge. The fund opened well, rising about 35 percent in 2010.
Then the trade broke. Even as the metal itself climbed to a record near 1,900 dollars an ounce in September 2011, the gold fund finished that year down close to 36 percent, dragged by mining stocks and leverage that fell far harder than bullion. It kept bleeding the next two years. When gold finally cracked into a bear market in 2013, the fund collapsed by roughly 60 percent. The damage was not confined to the gold vehicle. His most leveraged fund, Advantage Plus, lost about 51 percent in 2011, and the flagship Advantage fund lost roughly 36 percent the same year.
The Sino-Forest position
Sino-Forest Corporation was a Toronto-listed Chinese forestry company that Paulson & Co. held as one of the larger positions in its Advantage funds, the firm’s event-driven equity vehicles. The diligence record on the position, reconstructed afterward from Paulson’s own statements and the firm’s response to the short-seller report, looked nothing like the work that had gone into the 2007 mortgage trade. The team had not commissioned a database. They had not built a parallel set of records to verify what the company was telling them. There was no equivalent of Pellegrini’s eighteen-month data project. No one from the firm had walked the Yunnan timberland the company claimed to own. The position rested on audited financial statements from Ernst & Young, the company’s standing as a long-listed Toronto issuer, and the sell-side analyst reports that covered the stock.
On June 2, 2011, Muddy Waters Research, the short-selling firm run by Carson Block, published a report alleging that Sino-Forest had massively overstated its timber holdings in Yunnan province and that its accounting was, in the report’s terms, a fraud. Block’s team had done the walking work that Paulson’s had not, visiting regional administrative offices, pulling local filings, and documenting the gap between the plantations the company claimed and what could be found on the ground. The stock began to crater immediately.
Paulson’s team initially stood by the position, citing the audited financials and the long public track record. The defense did not last. Within two weeks the firm reversed course and, as a regulatory filing showed, disposed of its entire 34.7 million-share stake by June 17, a loss Bloomberg reported at roughly 720 million dollars. The stock lost more than 80 percent of its value over those weeks. Sino-Forest filed for bankruptcy protection in Canada on March 30, 2012. Five years later, the Ontario Securities Commission found that senior Sino-Forest executives had committed fraud, a finding later upheld on appeal. Block had been right.
The Sino-Forest loss was not even the largest dollar loss in the 2011 drawdown. Losses across Bank of America, Citigroup, and other bets on a U.S. recovery were comparable or larger over the year. But it was the most narratively damaging, because the gap between the two diligence processes was so stark. The team that had pulled county-level deed records to verify a thesis about American mortgages had bought a Chinese forestry company on the strength of an audit and a Toronto listing. The edge that produced the subprime trade was process. The Sino-Forest position showed what the firm looked like without it.
The drawdown years
Assets under management peaked around 38 billion dollars in 2011 and declined steadily from there. Outside investors redeemed. The Advantage funds posted further losses, the gold fund never recovered to its 2010 and 2011 highs, and by the middle of the decade firm assets had fallen sharply, a growing share of what remained being Paulson’s own money and that of his family and a small group of long-term partners.
In July 2020, Paulson announced he was converting the firm into a family office and returning outside capital. He had run an external hedge fund for 26 years. The conversion was framed as a choice rather than a forced retirement, but the trajectory of redemptions and performance suggested the decision had been pending for some time.
What the trade record actually shows
The clean version of the Paulson story, billion-dollar win, billion-dollar loss, retreat to a quiet office, is too neat. The 2007 trade was not a flash of inspiration but the output of eighteen months of database work, custom-structured synthetic deals with the dealer banks, and a willingness to pay insurance premiums on positions that lost money for over a year before paying off. The 2011 loss was not a single bad bet but a compounding of leveraged long positions in financial stocks during the European debt crisis, alongside the Sino-Forest blowup and a gold fund that broke even as the metal it tracked kept rising.
The through line is the diligence asymmetry. On subprime, the firm built the data the rest of the market did not have. On Sino-Forest, the firm relied on the data the rest of the market was already looking at. The first produced close to four billion dollars in personal compensation in a single year. The second produced a loss of hundreds of millions, sold off in a fortnight once someone who had actually walked the ground published what he found.
Paulson is 70 years old. He still runs money from Midtown, though the firm is a family office now, returning the outside capital it once fought to attract. The staff is a fraction of what it was at the peak, a handful of people where there used to be dozens. And gold is the conviction he never fully walked away from, the bet on an inflation that did not arrive the way he expected, carried forward long after the year it cost him so much.