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

Abstract

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.

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Improving neural networks by preventing co-adaptation of feature detectors
• On the difficulty of training Recurrent Neural Networks
• Res2Net: A New Multi-scale Backbone Architecture
• Gaussian Error Linear Units (GELUs)
• Towards a Definition of Disentangled Representations
• Concept Whitening for Interpretable Image Recognition
• Sparse Principal Component Analysis via Variable Projection
• Concept Bottleneck Models
• Deep Learning in Characteristics-Sorted Factor Models
• Continuously Differentiable Exponential Linear Units