Cross-Stock Predictability via LLM-Augmented Semantic Networks

TL;DR

This paper introduces a two-stage framework using large language models to filter and refine text-based financial networks, improving cross-stock return predictability and trading performance.

Contribution

It presents a novel LLM-augmented method for filtering economic relations in financial networks, enhancing their predictive power and economic relevance.

Findings

LLM-based filtering increased Sharpe ratio from 0.742 to 0.820.

Refined networks reduced maximum drawdown from 10.47% to 7.85%.

Improved cross-stock predictability in backtests.

Abstract

Text-based financial networks are increasingly used to study cross-stock return predictability. A common approach constructs links from similarities in firms' disclosure embeddings, but such networks often contain spurious edges because textual proximity does not necessarily imply economic connection. We propose a two-stage framework that first builds a sparse candidate graph from 10-K embeddings and then uses a large language model to classify and filter candidate edges according to their economic relations. The refined graph is used to aggregate pair-level mean-reversion signals into stock-level trading signals with relation-aware and distance-based weights. In a backtest on S&P 500 constituents from 2011 to 2019, LLM-based edge filtering improves the long-short Sharpe ratio from 0.742 to 0.820 and reduces maximum drawdown from $-$10.47% to $-$7.85%. These results suggest that LLM-based reasoning can improve the economic fidelity of text-derived financial networks and strengthen cross-stock predictability.