Which Voices Move Markets? Speaker Identity and the Cross-Section of Post-Earnings Returns

Paper: arXiv 2604.13260

Authors: Karmanpartap Singh Sidhu, Junyi Fan, Maryam Pishgar

Abstract

We utilize FinBERT, a domain-specific transformer model, to parse 6.5 million sentences from 16,428 S&P 500 quarterly earnings call transcripts (2015-2025) and demonstrate that post-earnings stock returns are not equally affected by all speakers in a conference call. Our section-weighted sentiment, with empirically derived speaker weights (Analyst 49%, CFO 30%, Executive 16%, Other 5%), achieves an out-of-sample Spearman IC of 0.142 versus 0.115 in-sample, generates monthly long-short alpha of 2.03% unexplained by the Fama-French five-factor model (t = 6.49), and remains significant after controlling for standardized unexpected earnings (SUE). FinBERT section-weighted sentiment entirely subsumes the Loughran-McDonald dictionary approach (FinBERT t = 5.90; LM t = 0.86 in the combined specification). Signal decay analysis and cumulative abnormal return charts confirm gradual price adjustment consistent with sluggish assimilation of soft information. All results undergo rigorous out-of-sample validation with an explicit temporal split, yielding improved rather than deteriorated predictive power.

Complexity vs Empirical Score

• Math Complexity: 3.5/10
• Empirical Rigor: 6.0/10 (capped from 8.5: transaction costs not modelled; see penalty flags)
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper presents strong empirical evidence using NLP models and robust econometric controls (Fama-French 5-factor, SUE) across a massive dataset, proving distinct alpha from speaker-weighted sentiment. While the mathematical complexity is relatively low, focusing on applied statistical regression and transformer embeddings, the empirical validation and out-of-sample rigor are top-tier.

Research Flowchart

  flowchart TD
    A["Research Question: Do specific speaker identities<br>in earnings calls drive post-earnings returns?"] --> B["Input Data: 16,428 S&P 500 Transcripts<br>(6.5M sentences, 2015-2025)"]
    B --> C["Processing: FinBERT Sentiment Parsing &<br>Empirical Speaker Weighting"]
    C --> D["Speaker Weights Derived:<br>Analyst (49%), CFO (30%), Exec (16%), Other (5%)"]
    D --> E["Out-of-Sample Validation &<br>Performance Analysis"]
    E --> F["Key Findings:<br>• Out-of-sample Spearman IC = 0.142<br>• Monthly Long-Short Alpha = 2.03% (t=6.49)<br>• Subsumes Loughran-McDonald Dictionary"]