2iQ’s Insider Data Underpins Research White Paper from RavenPack

2 minutes read
Page's meta description as image.

RavenPack, a leading data analytics provider in financial services, recently released a white paper exploring whether the public discourse of company insiders aligns with their own transactions as insiders.

Authored by Peter Hafez, Ricard Matas, Alan Liu and Marko Kangrga, RavenPack’s detailed analysis is titled Combining Earnings-Call Transcripts and Insider Transactions.

The RavenPack team used indicators developed using 2iQ’s insider transactions data to conduct their research, which led 2iQ’s own Head of Data Science, Quantitative Research to share her thoughts.

A compelling use of insider transactions to enhance market predictions.

As with many trading signals that are derived from alternative data feeds, there is often a lot of noise in the data that can interfere with our understanding. Developing systematic strategies in this space involves careful considerations and unique approaches, especially since we are usually dealing with only one type of information flow for a given dataset.

In combining an earnings transcripts sentiment score to an insider score (derived from 2iQ Insider Transactions data), RavenPack demonstrates how, when done carefully, overlaying strategies can effectively hone in on pertinent information to make a signal stronger.

The RavenPack aggregated score is constructed in a way that overweights stocks where insider trends and transcript sentiment align, and underweights stocks where they do not. Essentially betting more on companies whose words match the actions of their top insiders, and betting less on those that don’t.

The result is a strategy with a performance that is stronger and more long-lived and particularly across longer trading horizons, relative to the transcript-only strategy.

It is great to see the clear and positive impact our insiders dataset has on the original signal, and I believe there is a lot of scope to extend this overlay approach to other compatible datasets.

You can download RavenPack’s white paper here.