Man versus Machine Learning Revisited

Yingguang Zhang; Yandi Zhu; Juhani T. Linnainmaa

doi:10.1093/rfs/hhaf066

Man versus Machine Learning Revisited

Yingguang Zhang¹; Yandi Zhu²; Juhani T. Linnainmaa³

¹ Guanghua School of Management, Peking University · ² School of Finance, Central University of Finance and Economics · ³ Dartmouth College, NBER, and Kepos Capital ,

Review of Financial Studies 2025

Binsbergen, Han, and Lopez-Lira (2023) predict analysts’ forecast errors using a random forest model. A strategy that trades against this model’s predictions earns a monthly alpha of 1.54% ($ t $-value = 5.84). This estimate represents a large improvement over studies using classical statistical methods. We attribute the difference to a look-ahead bias. Removing the bias erases the alpha. Linear models yield as accurate forecasts and superior trading profits. Neither alternative machine learning models nor combinations thereof resurrect the predictability. We discuss the state of research into the term structure of analysts’ forecasts and its causal relationship with returns.

DOI: 10.1093/rfs/hhaf066
Volume: 38 (12)
Pages: 3768-3790
Language: en
Export: BibTeX
Sources: openalex crossref