Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Corporate Earnings Calls and Analyst Beliefs

Giuseppe Matera

TL;DR

The paper investigates how narratives embedded in corporate earnings calls shape financial analysts' beliefs and subsequent realized earnings. It combines a rich data suite of call transcripts, analyst forecasts, and firm fundamentals with machine-learning models to show that textual content adds predictive power beyond traditional quantitative predictors. A novel in-silico morphing framework uses LLM-generated counterfactual transcripts to quantify Predicted Treatment Effects (PTEs) of six narrative dimensions (Guidance, Jargon, Confidence, Global Focus, Sentiment, Uncertainty), revealing systematic over- and under-reactions by analysts to specific narratives. The approach demonstrates that language, even when not adding quantitative information, meaningfully informs expectations and can influence market dynamics, with potential implications for both corporate communication strategy and analyst judgment. The study also introduces an automated LLM-based judge to ensure credibility of narrative morphs, strengthening causal interpretation of narrative effects in financial markets.

Abstract

Economic behavior is shaped not only by quantitative information but also by the narratives through which such information is communicated and interpreted (Shiller, 2017). I show that narratives extracted from earnings calls significantly improve the prediction of both realized earnings and analyst expectations. To uncover the underlying mechanisms, I introduce a novel text-morphing methodology in which large language models generate counterfactual transcripts that systematically vary topical emphasis (the prevailing narrative) while holding quantitative content fixed. This framework allows me to precisely measure how analysts under- and over-react to specific narrative dimensions. The results reveal systematic biases: analysts over-react to sentiment (optimism) and under-react to narratives of risk and uncertainty. Overall, the analysis offers a granular perspective on the mechanisms of expectation formation through the competing narratives embedded in corporate communication.

Corporate Earnings Calls and Analyst Beliefs

TL;DR

The paper investigates how narratives embedded in corporate earnings calls shape financial analysts' beliefs and subsequent realized earnings. It combines a rich data suite of call transcripts, analyst forecasts, and firm fundamentals with machine-learning models to show that textual content adds predictive power beyond traditional quantitative predictors. A novel in-silico morphing framework uses LLM-generated counterfactual transcripts to quantify Predicted Treatment Effects (PTEs) of six narrative dimensions (Guidance, Jargon, Confidence, Global Focus, Sentiment, Uncertainty), revealing systematic over- and under-reactions by analysts to specific narratives. The approach demonstrates that language, even when not adding quantitative information, meaningfully informs expectations and can influence market dynamics, with potential implications for both corporate communication strategy and analyst judgment. The study also introduces an automated LLM-based judge to ensure credibility of narrative morphs, strengthening causal interpretation of narrative effects in financial markets.

Abstract

Economic behavior is shaped not only by quantitative information but also by the narratives through which such information is communicated and interpreted (Shiller, 2017). I show that narratives extracted from earnings calls significantly improve the prediction of both realized earnings and analyst expectations. To uncover the underlying mechanisms, I introduce a novel text-morphing methodology in which large language models generate counterfactual transcripts that systematically vary topical emphasis (the prevailing narrative) while holding quantitative content fixed. This framework allows me to precisely measure how analysts under- and over-react to specific narrative dimensions. The results reveal systematic biases: analysts over-react to sentiment (optimism) and under-react to narratives of risk and uncertainty. Overall, the analysis offers a granular perspective on the mechanisms of expectation formation through the competing narratives embedded in corporate communication.

Paper Structure

This paper contains 38 sections, 11 equations, 9 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Subjective beliefs and analyst expectations.
  3. LLMs as tools for empirical social science.
  4. Financial disclosures and textual analysis.
  5. Contribution.
  6. Outline.
  7. Data
  8. Corporate Earnings Calls
  9. Analyst Forecasts
  10. Matching Forecasts to Earnings Calls
  11. Accounting and Market-Based Information
  12. Jensen-Kelly-Pedersen (JKP) stock characteristics
  13. Standardized Unexpected Earnings (SUE).
  14. Financial Statements
  15. Methodology
  16. ...and 23 more sections

Figures (9)

  • Figure 1: Earnings Calls: Number per Year and Average Length
  • Figure 2: Timing of Analyst Forecasts Around Earnings Calls
  • Figure 3: Timeline of the Earnings Announcement Cycle
  • Figure 4: Term Structure of Textual Predictive Power
  • Figure 5: Predicted Effects of Fundamental Information
  • ...and 4 more figures