Ranking Stocks by Filing Lexical Density

Summary

This strategy uses language metrics from company filings to form a monthly long-short equity portfolio. It selects a stock universe using liquidity and market-cap filters, retrieves filing metrics, and ranks eligible companies separately on lexical density and specific density. For each metric, it buys the highest-ranked group and shorts the lowest-ranked group, then liquidates and rebalances. The sample implementation uses daily data, quarterly universe selection, and filings no older than thirty days at selection time.

The code describes portfolio construction and data handling, but the document gives no performance results, benchmark comparison, or evidence that denser filing language predicts returns. It also combines positions from the separate metric rankings and uses a custom fee assumption and leverage setting, which can materially affect outcomes. The example should therefore be read as an implementation sketch; it does not establish that the signal is profitable or robust.

Key ideas

• The strategy ranks stocks by lexical density and specific density from company filings.
• It forms long positions in the highest-ranked stocks and short positions in the lowest-ranked stocks.
• The universe is refreshed quarterly, while holdings are rebalanced monthly.
• The example filters stale filing observations and applies liquidity and market-cap screens.
• No backtest results are provided to validate the signal.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.