Measuring Equity Sentiment with Traditional and Alternative Data

Summary

This translated research review examines seven U.S. equity sentiment signals built from analyst revisions and surprises, short-selling data, news, options, consumer web content, corporate filings, and insider trades. It describes how the signals are standardized and combined into an equal-weight composite, then assessed as pure factors after controlling for established equity styles. The study also compares signal exposures with other styles and tests long-only portfolio constructions: an optimization approach that targets high factor exposure while controlling risk, and a simpler rule-based approach intended to accommodate greater capacity.

The reported historical analysis found positive pure-factor results for most measures, and the composite outperformed its components over the study period. Simulated portfolios often exceeded benchmark returns or Sharpe ratios, with exceptions including news, options, consumer, or insider signals depending on the test. These are historical simulations, not guarantees. Data coverage and history vary by signal, alternative sources may be limited, and results span a particular U.S. universe and period. Faster-changing signals may also require more frequent rebalancing and can increase turnover.

Key ideas

• The study constructs seven sentiment measures from both conventional financial data and alternative sources.\nIt standardizes signal exposures and forms a composite using equal weights among available signals.\nMost signals had positive historical pure-factor performance, while the composite performed better than individual components in the reported sample.\nLong-only tests compare risk-controlled optimization with a simpler, higher-capacity weighting rule.\nSignal history, coverage, turnover, and performance differ, limiting generalization beyond the studied period and universe.

Tags

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