Building an Insider Trading Database and Predicting Future Equity Returns

40 Pages Posted: 18 Jul 2017  

John Ryle

Northwestern University, Predictive Analytics, Students

Date Written: July 11, 2017

Abstract

This paper considers the impact of legal insider trading on subsequent investment returns. With filings data gathered directly from the Security Exchange Commission, various machine learning algorithms are applied, scored and compared for their return predictability relative to the Russell 3000 index. Predictor variables are derived directly from the insider filings, as well as from metadata from related corporate filings. The filings data, spanning years 2005 through 2015, is transformed into a relational database and analyzed using statistical analysis and predictive modeling libraries within the R programming language. The data is separated into training, validation and test sets. Modeling was performed using a variety of algorithms. Performance in the validation set appeared quite predictive, but results were less conclusive when scored on the test set.

Keywords: insider trading, SEC Form, $ machine learning

Suggested Citation

Ryle, John, Building an Insider Trading Database and Predicting Future Equity Returns (July 11, 2017). Available at SSRN: https://ssrn.com/abstract=3000704 or http://dx.doi.org/10.2139/ssrn.3000704

John Ryle (Contact Author)

Northwestern University, Predictive Analytics, Students ( email )

Evanton, IL
United States

Register to save articles to
your library

Register

Paper statistics

Downloads
146
rank
181,545
Abstract Views
2,129
PlumX