Relative Sentiment and Machine Learning for Tactical Asset Allocation

Relative Sentiment and Machine Learning for Tactical Asset Allocation

AuthorsRaymond Micaletti

Paper linkhttps://papers.ssrn.com/sol3/papers.cfm?abstract_id=3475258

Uses Sentix 6-month economic expectation indices to construct relative sentiment factors for the U.S., Europe, Japan, and Asia ex-Japan. Machine learning forecasts one-month forward equity returns across regions.

Relative sentiment factors capture predictive signals in equity markets over intermediate horizons, with significant results across regions, corroborating prior evidence that institutional versus private sentiment differences affect returns.