Machine Learning Nowcasting of Price-Earnings Ratios

A term originated in meteorology, nowcasting pertains to the prediction of the present and very near future. Nowcasting applications in economics and finance are intrinsically a mixed frequency data problem as the object of interest is a low-frequency data series (e.g., quarterly), whereas the information is real-time high frequency (e.g., daily, weekly, or monthly). Eric Ghysels, in collaboration with Andrii Babii (UNC Chapel Hill), Ryan Ball (University of Michigan) and Jonas Striaukas (UC Louvain), studies nowcasting of price-earning ratios, which are quarterly, using monthly and daily information flows. They propose new methods, namely machine learning MIDAS panel regression models and show that for a significant number of firms in their sample nowcasting can be fully automated whereas for the remaining set there is a potential to improve upon analyst forecasts by judicious combination of their insights with machine learning algorithms.

Watch Webinar Read on arXiv

Eric Ghysels