Interpretable Deep Learning for Stock Returns: A Consensus-Bottleneck Asset Pricing Model

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Changeun, Jeong, Younwoo, Jang, Bong-Gyu
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911620690608128
author Kim, Changeun
Jeong, Younwoo
Jang, Bong-Gyu
author_facet Kim, Changeun
Jeong, Younwoo
Jang, Bong-Gyu
contents We introduce the Consensus-Bottleneck Asset Pricing Model (CB-APM), which embeds aggregate analyst consensus as a structural bottleneck, treating professional beliefs as a sufficient statistic for the market's high-dimensional information set. Unlike post-hoc explainability approaches, CB-APM achieves interpretability-by-design: the bottleneck constraint functions as an endogenous regularizer that simultaneously improves out-of-sample predictive accuracy and anchors inference to economically interpretable drivers. Portfolios sorted on CB-APM forecasts exhibit a strong monotonic return gradient, robust across macroeconomic regimes. Pricing diagnostics further reveal that the learned consensus encodes priced variation not spanned by canonical factor models, identifying belief-driven risk heterogeneity that standard linear frameworks systematically miss.
format Preprint
id arxiv_https___arxiv_org_abs_2512_16251
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Interpretable Deep Learning for Stock Returns: A Consensus-Bottleneck Asset Pricing Model
Kim, Changeun
Jeong, Younwoo
Jang, Bong-Gyu
Pricing of Securities
Artificial Intelligence
Machine Learning
We introduce the Consensus-Bottleneck Asset Pricing Model (CB-APM), which embeds aggregate analyst consensus as a structural bottleneck, treating professional beliefs as a sufficient statistic for the market's high-dimensional information set. Unlike post-hoc explainability approaches, CB-APM achieves interpretability-by-design: the bottleneck constraint functions as an endogenous regularizer that simultaneously improves out-of-sample predictive accuracy and anchors inference to economically interpretable drivers. Portfolios sorted on CB-APM forecasts exhibit a strong monotonic return gradient, robust across macroeconomic regimes. Pricing diagnostics further reveal that the learned consensus encodes priced variation not spanned by canonical factor models, identifying belief-driven risk heterogeneity that standard linear frameworks systematically miss.
title Interpretable Deep Learning for Stock Returns: A Consensus-Bottleneck Asset Pricing Model
topic Pricing of Securities
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2512.16251