Hedge Fund Leverage Crisis & Quant Opportunities

Today’s Picks

IMF Warns of Hedge Fund Leverage Risk: Scale Tripled in a Decade

Hedge funds have become significant participants in the US Treasury market, increasing their holdings from 4% in 2022 to approximately 9% today.

This expansion relies heavily on leverage, including synthetic leverage constructed via derivatives. While this aids liquidity and price discovery in normal environments, it can amplify volatility and trigger market dislocations during periods of stress.¹

Industry News

Eisler Capital Spends Nearly $500M on Staff Before Shutdown

Eisler Capital, a multi-strategy hedge fund founded by former Goldman Sachs partners, managed approximately $4 billion in assets.

It adopted a “pass-through” fee model, requiring investors to directly bear all fund compensation costs.

To poach talent from giants like Citadel, its staff costs surged by over 900% in five years. In its final year before shutdown, average per-employee compensation remained at $1.5 million, totaling nearly $444 million. However, these high costs did not translate into performance: Eisler posted a full-year loss of 14.3% in 2025 before ultimately liquidating.

Eisler Capital’s closure was not a “generous severance” but a suicide of its business model. When talent costs far exceed trading returns, larger scale accelerates losses. Eisler’s lesson is clear: without top-tier performance, mimicking a giant’s high-salary pass-through model only hastens demise. ²

Quant Research

Investing Is Compression: Oscar Stiffelman’s Latest View

Computer scientist and early Google member Oscar Stiffelman, founder of Nand Capital (an information-theory hedge fund), recently published the 4th edition of Investing Is Compression on arXiv. Below is a summary.

In 1956, John Kelly wrote a paper at Bell Labs describing the relationship between gambling and information theory. The subsequent Kelly Criterion is regarded as an objective, closed-form solution for determining bet size given known odds and margins.

Samuelson argued it was arbitrary and subjective, successfully excluding it from mainstream economics. Fortunately, it survived in computer science, largely due to Tom Cover’s work at Stanford. He demonstrated it to be the uniquely optimal method for investing: maximizing long-term wealth, minimizing bankruptcy risk, and being competitively optimal in a game-theoretic sense, even in the short term.

One of Cover’s most surprising contributions to portfolio theory is the “Universal Portfolio.” Related to universal compression in information theory, it performs asymptotically as well as the best constant-rebalanced portfolio in hindsight.

I borrow a trick from this algorithm to show that Kelly’s objective, even in its general form, decomposes the investing problem into three parts: a capital term, an entropy term, and a divergence term. The only way to maximize growth is to minimize divergence, which measures the difference between our distribution and the true distribution in bits. Investing is fundamentally a compression problem.

This decomposition also yields new practical results. Since the capital and entropy terms are constant across strategies in a given backtest, the difference in log growth between two strategies measures their relative divergence in bits.

I also introduce a win-rate heuristic that allocates capital based on the dominant probability of each asset in the candidate set. The growth gap relative to the optimal portfolio is bounded by the entropy bound of the win-rate distribution. To my knowledge, both this heuristic and the entropy bound are original contributions. ⁴

Shorting Seven Asset Classes on the Day After US Election: 1.86% Annualized Return, Sharpe 0.31

Before major elections, asset prices often include an additional premium driven by “political uncertainty.”

The question is: when and how does this premium dissipate across asset classes once the outcome is clear? This article from Quantpedia investigates this issue.

The paper notes that investors pay a high “political flood insurance” premium before voting, which is liquidated at the D+1 open.

In terms of strategy, an equal-weight short position in the D+1 risk asset portfolio yields an annualized return of 1.86%, a Sharpe ratio of 0.31, and a max drawdown of only -4.16%.⁵

Quant Career

Who Becomes a Point72 Intern?

Point72 is a global multi-strategy hedge fund giant led by Steve Cohen, managing approximately $58.5 billion in assets.

In a blog post on October 7, Point72 detailed the daily routines of three summer interns in Tokyo, Warsaw, and Hong Kong, offering direct reference value for applicants.

The Tokyo Academy intern rotates with Portfolio Managers (PMs). Daily tasks include tracking overnight markets, attending investment research meetings, and conducting deep-dive research on designated companies, including building financial models, performing due diligence, validating investment theses, and reporting daily to PMs. This role requires a foundation in fundamental research and modeling, as well as the ability to prioritize amidst information overload and balance depth with efficiency.

The Warsaw Infrastructure Automation intern becomes familiar with internal applications and workflows. The core responsibility is optimizing an AI tool that automatically handles routine tasks. This role suits candidates with a computer science background, understanding of automation processes, and the ability to learn quickly and collaborate remotely with global teams.

The Hong Kong Equity Management intern tracks market data, liaises with PMs, analysts, and trading/quant teams, and leads the development of a dashboard tracking compliant alternative data, which is already used in live trading meetings. This role requires data processing and visualization skills, cross-team communication abilities, and an entrepreneurial spirit to take initiative.

Based on their feedback, all three roles test cross-team communication and the ability to drive progress independently under guidance. ⁶