---
title: Contextualizing Corporate Governance with AI Proxy Advice: Evidence from Proxy Contests
url: https://www.ml-quant.com/papers/ssrn/7683760/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-11-14
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: SSRN 7683760
source_url: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7683760
featured: 2026-11-13
citations: unknown
topic: ML & AI Methods
---


# Contextualizing Corporate Governance with AI Proxy Advice: Evidence from Proxy Contests

An AI-based proxy advisor using firm context and text embeddings beats traditional advisors by 6 to 7 percentage points, earning 10.3% annualized alpha on contested board decisions.

- Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7683760
- Identifier: SSRN 7683760
- Released: 2026-11-12
- First featured: Quant Letter No. 139 (2026-11-13): https://www.ml-quant.com/issues/2026-11-13/
- Citations (Semantic Scholar): not tracked
- Published in: not yet
- Topic: ML & AI Methods
- Authors: Sean S. Cao, Wenzhi Ding, Tao Li, Wei Jiang

## Related

- [STOCK-JEPA: Prior-Anchored Latent Revision Representation Learning in Equity Markets](https://www.ml-quant.com/papers/arxiv/2610.07006/): A joint-embedding framework learns predictable incremental revisions from market data relative to a financial prior, combining interpretability from classic models with the pattern-capture strength of deep learning.
- [NMIXX: Domain-Adapted Neural Embeddings for Cross-Lingual eXploration of Finance](https://www.ml-quant.com/papers/arxiv/2507.09601/): A Meta-Analysis: The NMIXX model, fine-tuned with high-confidence triplets, excels in capturing financial semantics in low-resource languages like Korean, emphasizing the importance of tokenizer design in cross-lingual settings.
- [NeuralFactors: A Novel Factor Learning Approach to Generative Modeling of Equities](https://www.ml-quant.com/papers/arxiv/2408.01499/): The research presents NeuralFactors, a machine-learning method for factor analysis that improves performance and efficiency in stock embedding.
- [Market-Conditioned JEPA: Learning from Unpredictable Market Shocks](https://www.ml-quant.com/papers/arxiv/2611.13572/): The study conditions self-supervised learning on market factors to improve prediction of returns, exposure, risk, and liquidity across minute-level stock data.
- [Exact Phase-Type Point Processes for Certified Measurement and Prediction of Event Streams](https://www.ml-quant.com/papers/arxiv/2610.14948/): A multiscale phase-type Hawkes model measures how fast prices absorb macroeconomic news versus self-exciting echoes, showing direct absorption in about one second and extended responses up to 137 seconds.
- [Deep kernel hedging](https://www.ml-quant.com/papers/arxiv/2609.34474/): A hybrid approach combining neural networks with kernel methods produces robust derivative hedges in low-data regimes and scales via random Fourier features with convergence guarantees.