Overview and details of the sessions of this conference. Please select a date or room to show only sessions at that day or location.
Please select "Show Presentations" and "List View" or a single session for detailed view (with abstracts, discussants, downloads). PDF downloads are accessible to registered participants only (after Login).
Analyst vs. machine learning: differences in implied cost of capital estimations
Minghui Chen
Technical University of Munich, Germany
This paper enhances the implied cost of capital (ICC) estimation by improving analysts' multi-horizon earnings forecasts with machine learning, which demonstrates reduced bias and higher forecast accuracy. The resulting machine learning model-based ICC is approximately 17% lower than the traditional analyst-based ICC with optimization bias. The model-based ICC shows stronger cross-sectional relationships with future realized returns, thus providing a better proxy for ex-ante expected returns. In addition, stocks with underestimated ICCs exhibit significant outperformance relative to those with overestimated ICCs. Weaker information environments primarily drive the ICC differences between analyst-based and model-based ICCs. Moreover, greater ICC differences are associated with lower earnings announcement returns.
