University of Denver
Utilizing Machine Learning to Forecast Sector-Level Equity Returns from Sector-Level Financial Ratios
Abstract
dc:description.abstractAcademics and equity analysts often utilize fundamental financial ratios to evaluate an industry's financial and operational performance. However, sophisticated econometric methods could be improved for estimating these ratios' importance in predicting sector-level stock return performance. To address this gap, gradient-boosted decision tree methods were applied to the WRDS sector-level financial ratio database to identify if financial ratios effectively forecast sector-level stock returns in future periods. The findings show that these ratios are predictive and economically material, both stand alone and in combination with other macroeconomic factors. Using these predictions, long and long-short sector-level portfolios were created, and most of these portfolios consistently outperformed the market in absolute and risk-adjusted returns. Furthermore, these portfolios generate positive alpha in the presence of the well-established Fama-French factor models. This research demonstrates that sector-level information is effective in forecasting equity market performance.
Degree
thesis:*- Name thesis:degree_name
- Ph.D.
- Level thesis:degree_level
- Dissertation
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kuppenheimer, Greg
- Contributors dc:contributor
-
- Jack Strauss
- Chris GauthierDickey
- Alex Petkevich
- Shahram Amini
Subjects
dc:subject × 13Rights
dc:rights- Statement dc:rights
-
- <p>Copyright is held by the author. User is responsible for all copyright compliance.</p>
- Greg Kuppenheimer
- Language dc:language
- English (eng)
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.du.edu/etd/2451