Bubbles are easy to identify in hindsight and extraordinarily difficult to profit from in real time. The subprime mortgage crisis is often cited as proof that savvy investors can, and should, short obvious excesses. But a closer look tells a different story. Even investors who correctly identified the housing bubble faced severe risks: timing uncertainty, negative carry, illiquidity, counterparty risk, and political intervention. The lesson isn’t that bubbles don’t exist. It’s that shorting them is usually the worst way to express that belief. Excerpts from Aaron Brown and Richard Dewey’s paper: Toil and Trouble, Don’t Get Burned Shorting Bubbles.
The Big Short
In 2010, Michael Lewis published The Big Short, a gripping account of the subprime mortgage crisis and the small group of investors who bet against it. But with a decade of hindsight, we now see that while a handful of traders became legends for betting against housing, many of them have struggled to replicate that success in the years since. Indeed, betting directly against a bubble is often a poor risk-reward proposition, especially when the bubble lacks a clear catalyst for correction.
Take today’s debates about an “AI bubble”. Even if you believe certain assets are overpriced, shorting them means grappling with timing, path, and intervention risk. Just as banks were rescued during the financial crisis, it’s not hard to imagine datacenter capex eventually being burdened by taxpayers in one way (China already heavily sponsors AI development). The difference between having a view and making money is structuring a position that survives the path dependency and still pays in the state of the world you’re betting on.
Credit vs Equity
One reason the subprime trade gravitated toward credit (buying CDS’ on MBS’, CDOs, and Synthetic CDOs) rather than equities is structural. Credit has a capped upside. Unlike equity, which can continue rising as long as the narrative holds, a bond can only trade up to the risk-free yield. If the thesis is that cash flows won’t support the capital structure, credit offers a more precise way to express that view. Indeed, very few investors were willing to short bank equities outright. The upside risk felt unbounded, timing was uncertain, and the political risk was obvious. But the subprime experience also shows the limits of that logic. While credit shorts avoid some of the path risk of equities, they introduce others.
Timing risk remains. Early shorts were effectively betting against older mortgages with relatively low default rates, while later mortgages turned out to be far worse. Being right too early meant paying carry for years, with no guarantee the trade wouldn’t be better later. Policy risk also remained. Investors had to consider whether government intervention might support homeowners, banks, or both, potentially reducing or eliminating payouts. Political incentives favored stabilization over market discipline, adding another layer of uncertainty to the trade.
Finally, structure and counterparty risk proved critical. CDS were illiquid, bilateral instruments dependent on dealer marks rather than transparent market prices. Even when the thesis was clearly working, adverse marks triggered margin calls, delaying or jeopardizing payouts. Investors also faced the risk that their counterparties might fail precisely when they were owed the most.
Alternative Trades
Basis Trade: Magnetar reportedly employed a strategy whereby they purchased the riskiest equity tranche in many CDOs which often offered double-digit returns. They used this positive carry to pay for protection on the AAA tranches that most investors assumed were safe. Magnetar appreciated that the correlation between the safest AAA tranche and the lowest quality equity tranche would be close to one in a crisis due to the way these securities were constructed. This structure gave the trade a positive carry profile, eliminating the worry over timing. It also reduced the risk of government intervention as a bailout that eliminated payouts on the AAA protection would likely have produced gains, or at least prevented losses, in the equity tranches.
Buy the Dip: George Soros and his Chief Investment Officer Keith Anderson smelled opportunity and hired two ex-Salomon Brothers mortgage experts, Mason Haupt and Howie Rubin. Over the course of the next five years the Haupt-Rubin team reportedly made returns in excess of 30% annualized on roughly a billion dollars of capital.
Across the country in Newport Beach, the mortgage team at PIMCO, led by Dan Ivascyn seized on the same idea. Ivascyn bet heavily on subprime, by some accounts allocating over 50% of his Income Fund to the once shunned securities. From 2009 through 2018 the Income Fund grew from $223 million to $121 billion, with annualized performance of roughly 12%. That track record and asset growth helped Ivascyn gain power and eventually become CIO of PIMCO in 2015.
For those that didn’t want to deal with the complexity of subprime mortgages, buying the stock of large commercial banks offered an alternative on the same theme. David Tepper purchased shares of Citi and Bank of America near the bottom, helping his Appaloosa fund return 120% in 2009 on $12 billion in capital.
Cheap Convexity: Ray Dalio at Bridgewater and Jeff Talpins at Element Capital is rumored to have purchased futures or options on government bonds that would rise in value if the Fed aggressively cut interest rates. Many on Wall Street believe that Element purchased Eurodollar options in 2007 that helped his firm generate returns of 26.4% in 2007 and 34.9% in 2008. Talpins has posted annualized returns north of 20%, without a single losing year in the decade that followed. And simply being long volatility in equity markets, fixed income or currency markets paid off nicely for many traders.
What distinguished the most successful trades wasn’t superior foresight, but better structure. Investors like George Soros, David Tepper, and the mortgage teams at PIMCO didn’t rely on heroic timing. They put capital to work in trades that could scale and survive. It’s also worth noting that some alternatives offered far better payoff profiles with even lower carry. In hindsight, shorting specific institutions proved far more efficient than shorting housing broadly. Credit protection on Lehman, Bear Stearns, and AIG traded at remarkably low spreads before the crisis, offering significantly higher returns per unit of capital than subprime CDS.
Takeaways on Trading the “AI Bubble”
Betting against subprime mortgages ultimately worked, but it was a narrow, almost Goldilocks trade. Defaults had to rise enough to trigger insurance payouts, yet not so much that the financial system collapsed or the government intervened in ways that erased the trade. That balance was impossible to know in advance. Being “right” required not just a correct view, but a very specific path through the crisis.
That lesson matters for how we think about today’s AI boom. Few people dispute that AI is transformational. The real question is how much of the value it creates will actually be captured by investors, and by which part of the capital structure. As William Nordhaus has shown, the creators of new technologies historically capture only a small fraction of the total economic value, perhaps as little as 2%. The remaining surplus flows through to society in the form of lower prices, higher productivity, and new capabilities.
As private credit moves deeper into AI infrastructure, investors now have access across the entire capital structure. That broader toolkit allows views on timing, durability, and value capture to be expressed. This matters in a world where policy intervention is no longer exceptional. Since 2008, unconventional monetary policy and large fiscal responses have become routine, making clean, fast crises harder to bet on and increasing the value of trades that can withstand delay and intervention. The lesson from subprime isn’t that bubbles are impossible to spot, but that profiting from them depends on structure. Time horizon, liquidity, sizing, and carry often matter more than being directionally correct. The best opportunities are often found not by shorting the bubble itself, but by trading around it.
Disclaimer
The information provided on TheLogbook (the “Substack”) is strictly for informational and educational purposes only and should not be considered as investment or financial advice. The author is not a licensed financial advisor or tax professional and is not offering any professional services through this Substack. Investing in financial markets involves substantial risk, including possible loss of principal. Past performance is not indicative of future results. The author makes no representations or warranties about the completeness, accuracy, reliability, suitability, or availability of the information provided.
This Substack may contain links to external websites not affiliated with the author, and the accuracy of information on these sites is not guaranteed. Nothing contained in this Substack constitutes a solicitation, recommendation, endorsement, or offer to buy or sell any securities or other financial instruments. Always seek the advice of a qualified financial advisor before making any investment decisions.