Testing Analyst Recommendations as Short-Horizon Trading Signals

Summary

The document considers whether consensus analyst recommendations and revisions can contribute to trading signals, particularly for holding periods of hours to a few days. It notes the author’s impression from academic research that recommendations may contain little standalone information, and raises the possibility that they are more suited to longer-term strategies. Proposed initial analysis includes measuring correlations between analyst publications and price movements using rank-based or linear correlation methods.

A further hypothesis is that price moves may occur before analyst announcements, potentially reflecting information leakage. The author suggests examining whether such pre-publication movements could inform a mean-reversion entry. These are research ideas and questions, not established findings: the document gives no data, test design, empirical results, or evidence that pre-announcement moves reverse. Any investigation would need to account for announcement timing, overlapping observations, and the difference between correlation and a tradable signal.

Key ideas

• Consensus analyst recommendations and revisions may be evaluated as additions to technical signals.
• The author questions whether analyst data has useful information for short holding periods.
• Correlation tests between publication data and prices are proposed as an initial analysis.
• Pre-announcement price moves are raised as a possible basis for testing mean reversion.
• The document presents hypotheses rather than empirical evidence of a profitable effect.

Tags

# Correlation between fundamental and market data


# Correlation between fundamental and market data












I got hold of a data set which contains fundamental data like analyst recommendations/revisions (consensus only) and I am trying to come up with an idea of how this could be used as a trading signal or as an addition to one of my other technical trading signals. From academic research that I have read, it sounds like there is very little informational content in such analyst recs as a stand-alone signal. My holding periods are anywhere between hours to 1-3 days, so I guess it might be better suited for buy and hold long-term strategies in the literature.

My initial idea would be to see if there is generally correlation between analyst publications and price movements (simple spearman/Pearson). The other idea that I got from some papers was, that the biggest correlations to prices happens $t-1$, due to insider information leakage. If I could see bigger moves happen before announcements, then one could leverage this as an entry point for a mean-reversion type strategy. If anyone has any idea how something like this could be done or any other idea about how to use the fundamental data that would be appreciated.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

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