---
title: Design choices, machine learning, and the cross-section of stock returns
url: https://www.ml-quant.com/papers/ssrn/5031755/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: SSRN 5031755
source_url: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5031755
featured: 2024-11-27
citations: 5
topic: Asset Pricing & Factors
---


# Design choices, machine learning, and the cross-section of stock returns

The performance of machine learning models in predicting stock returns is greatly influenced by their design choices, with nonstandard errors in portfolio returns surpassing standard errors by 59%.

- Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5031755
- Identifier: SSRN 5031755
- Released: 2024-11-23
- First featured: Quant Letter No. 76 (2024-11-27): https://www.ml-quant.com/issues/2024-11-27/
- Citations (Semantic Scholar): 5
- Published in: not yet
- Topic: Asset Pricing & Factors

## Related

- [The relationship between general equilibrium models with infinitely-lived agents and overlapping generations models, and some applications](https://www.ml-quant.com/papers/arxiv/2411.07674/): The paper establishes that a two-cycle equilibrium in a model with infinitely-lived agents can also exist in an overlapping generations (OLG) model, indicating that both models can experience equilibrium indeterminacy and rational asset price bubbles.
- [Unveiling True Talent: The Soccer Factor Model for Skill Evaluation](https://www.ml-quant.com/papers/arxiv/2412.05911/): The Soccer Factor Model (SFM) is a new method for evaluating soccer players' performance independently from their team's influence, allowing for more accurate comparisons.
- [NUMOSIM: A Synthetic Mobility Dataset with Anomaly Detection Benchmarks](https://www.ml-quant.com/papers/arxiv/2409.03024/): A Synthetic Mobility Dataset: The paper presents NUMOSIM, a synthetic mobility dataset for testing anomaly detection techniques, simulating realistic mobility scenarios and anomalies to improve geospatial mobility analysis.
- [Fundamental Properties of Linear Factor Models](https://www.ml-quant.com/papers/arxiv/2409.02521/): The article examines conditional linear factor models in asset pricing. It explores the relationships and characteristics of returns and factors using conditional means and covariances. This research lays the groundwork for defining and estimating these models.
- [The Cross-Section of Corporate Bond Returns](https://www.ml-quant.com/papers/ssrn/5130312/): A study of U.S. corporate bonds identifies four factors that provide strong return premiums after costs.
- [Dynamic Factor Correlation Model](https://www.ml-quant.com/papers/arxiv/2503.01080/): A new model with a unique variation-free parameterization of factor loadings shows adaptability and scalability in both small and large asset return environments.