Alternative Data

Overview

Alternative data is evidence about markets that does not come from price and volume alone — search activity, text or other observable behavior — and it carries a validation burden at least as strict as any price-based signal before it is treated as informative.

Core questions

• Does a novel data source carry information not already reflected in price?
• How should a new signal source be tested against the same overfitting risks as any price-based signal?

Mathematical formulation

Attention-proxy predictive regression

A simple predictive regression of next-period returns on the change in a search-volume index (SVI), the form used to test whether retail attention, measured indirectly through search activity, precedes price movement — the question any alternative-data source must answer before it is treated as a signal.

Methods we use

• Search-attention data as a market variable

  Da, Z., Engelberg, J., & Gao, P. (2011). In search of attention. Journal of Finance, 66(5), 1461–1499.

• Multiple-testing-aware validation of new signal sources

  Harvey, C. R., Liu, Y., & Zhu, H. (2016). …and the cross-section of expected returns. Review of Financial Studies, 29(1), 5–68; and Bailey, D. H., & López de Prado, M. (2014). The deflated Sharpe ratio. Journal of Portfolio Management, 40(5), 94–107.

Open problems

• How should a new data source's limited history be treated when it did not exist for the full length of an otherwise longer backtest window?
• What false-discovery correction is appropriate when many alternative-data sources are screened for signal in parallel?