What do Hedge Funds Think? Textual Analysis of Investor Letters
People
(Responsible)
Abstract
Hedge funds play a critical role in the asset management industry due to their use of complex strategies and the flexibility to operate with fewer restrictions than other institutional investors. They also move significant amounts of capital, influencing market liquidity and prices. Understanding the factors driving their performance and investment decisions is important for investors, researchers, and policymakers. However, the literature has faced hurdles in examining these due to the low disclosure requirements for hedge funds and the discretion fund managers have in revealing their strategies. The project addresses these gaps by utilizing a large corpus of hedge fund letters to investors spanning multiple decades. Some letters pre-date publicly available databases.
Existing literature has only been able to infer hedge fund managers' views indirectly from market data. In this project, we overcome this limitation by extracting their perspectives directly from their own words. Our analysis draws on investor letters from a representative sample of long-short equity and global macro funds, both active and defunct, to avoid survivorship bias. We identify four topics that are crucial for understanding hedge fund performance with significant societal value: macroeconomic expectations, stock market expectations, approaches to sustainable investing, and communication strategies used to explain performance to investors.
Macroeconomic Expectations: Hedge fund managers are sophisticated investors, and their expectations of macroeconomic variables are crucial for asset pricing and the transmission of economic and monetary policies. The project will investigate whether hedge fund managers’ expectations are more accurate than those of professional forecasters, particularly after public announcements, and how their views reflect their interpretation of policy communications.
Stock Performance Expectations: Hedge fund managers often discuss individual stocks in their letters. This research will assess whether fund managers can predict future stock performance better than other professional forecasters, such as analysts, and will explore the events and analyses that cause managers to change their views on stocks.
Sustainable Investing: The project will examine hedge fund managers’ attitudes toward sustainable investing and whether they exploit opportunities created by green investors divesting from certain assets. It will explore whether hedge funds view climate change as a risk to be hedged or a speculative opportunity, contributing to either the reduction or the maintenance of climate-related risk premia.
Communication Strategies and Trading Styles: Hedge fund managers’ explanations of their past performance in investor letters can reveal their investment styles and risk-taking behavior. The project will explore how managers justify underperformance or overperformance and whether their communication strategies affect their clients’ decisions, such as preventing capital redemptions or causing panic-driven withdrawals.
To address these questions, we will employ state-of-the-art NLP techniques to analyze the content of the investor letters. Our analyses will encompass sentiment analysis, unsupervised topic modeling, and keyword discovery drawing on the BERT language architecture. For example, to extract macroeconomic expectations, we identify forward-looking portions of the text referring to macroeconomic policies and then measure the sentiment using FinBERT.
In conclusion, this research aims to bridge existing gaps in comprehending hedge fund strategies and behavior by providing direct insights from managers' own communications. The findings will deepen our understanding of hedge funds' influence on financial markets and the economy, shaping perceptions of their role in areas such as market stability, policy transmission, and sustainable investing.