Analyzing Speech to Detect Financial Misreporting

////Analyzing Speech to Detect Financial Misreporting

Analyzing Speech to Detect Financial Misreporting

By | 2017-01-30T16:42:17+00:00 July 18th, 2012|Research Insights, Behavioral Finance|1 Comment
Print Friendly, PDF & Email
(Last Updated On: January 30, 2017)

Analyzing Speech to Detect Financial Misreporting

  • Jessen L. Hobson, William J. Mayew, Mohan Venkatachalam
  • A version of the paper can be found here.

Abstract:

“We examine whether vocal markers of cognitive dissonance are useful for detecting financial misreporting. We use speech samples of CEOs during earnings conference calls and generate vocal dissonance markers using automated vocal emotion analysis software. We begin by assessing construct validity for the software-generated dissonance markers by correlating them with four dissonance-from-misreporting proxies obtained in a laboratory setting. We find a positive association between these proxies and vocal dissonance markers generated by the software, suggesting the software’s dissonance markers have construct validity. Applying the software to CEO speech, we find that vocal dissonance markers are positively associated with the likelihood of irregularity restatements. The diagnostic accuracy levels are 11% better than chance and of similar magnitude to models based solely on financial accounting information. Moreover, the association between vocal dissonance markers and irregularity restatements holds even after controlling for financial accounting and linguistic based predictors. Our results provide new evidence on the role of vocal cues in detecting financial misreporting.”

Data Sources:

Laboratory Generated Data using Fifty-nine undergraduate volunteers from two large public U.S. universities and 1,572 conference calls available on Thomson Reuters Street Events.

Discussion:

Detecting deceptive financial reporting is important to investors, auditors, analysts and regulators. To developing a framework to predict deceptive financial reporting in a timely manner, this study tests a hypothesis: The probability of financial misreporting is positively associated with the extent of cognitive dissonance markers contained in the CEO’s voice.

At first, the authors generate a sample of truth-telling and misreporting subjects in a laboratory setting and invoke cognitive dissonance in the misreporting subjects. Then thet find the LVA (Layered Voice Analysis) dissonance markers extracted from subject interviews associated with four dissonance-from-misreporting proxies: belief revisions, confessions of misreporting, unexpected reported scores, and a factor score that combines these three individual proxies.

After demonstrating that voice based dissonance markers generated by LVA have validity, they explore the following model:

Pr (Misreporting) = f (Vocal Dissonance Markers, Dissonance Drivers Unrelated to Misreporting, Financial Statement Based Predictors of Misreporting, CEO Characteristics, Monitoring).

In words, the probability of misreporting is associated with a collection of indicators that suggest something doesn’t smell right.

The authors then measure cognitive dissonance using speech samples of CEOs from their interactions with analysts and investors during earnings conference calls. Logistic regressions reveal a positive association between vocal dissonance markers and adverse irregularity restatements. Therefore, cognitive dissonance in CEO speech can predict whether a firm’s quarterly financial reports will be adversely restated at better than random chance levels. The overall evidence suggests vocal cognitive dissonance markers can predict financial misreporting.

Investment Strategy:

  1. Exam quarterly earnings call to detect fraud risk.
  2. Buy stocks with no or little fraud risk.
  3. Short stocks with high fraud risk
  4. Rebalance quarterly.
  5. Make money.

Commentary:

This paper provides evidence that elements of voiced speech can help detect financial misreporting. This paper is the first to investigate the predictive ability of vocal cues for actual misreporting in a capital market setting.
This research is informative to investors, analysts, and auditors who attempt to use earnings conference calls as an information source for assessing the risk of misreporting.
Vocal features represent only one class of deception markers and an evaluation of alternative nonverbal markers of deception is a worthwhile endeavor. Future research or investment could consider both verbal and nonverbal markers of deception, not only CEO’s speech but also CFO’s speech as well to detect misreporting and get a better view of all related information.

Go and there and hire you some linguistic PhDs (akin to what RenTec was doing 20 years ago) and try to make some money!


  • The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
  • Join thousands of other readers and subscribe to our blog.
  • This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.

About the Author:

Prior to joining the Alpha Architect team, Ms. Yao was a Research Assistant to Dr. Gray. She studied quantitative models and summarized over 200 academic articles on psychology and behavioral finance. Her prior experience includes work as a financial analyst at United Asset Growth (China) LLC, and as a business development intern for Shanghai Pudong Development Bank. Tian earned a Masters in Finance at Drexel University. She earned her Bachelors degree in Finance at Nanjing Normal University, China.