The First Sign: Detecting Future Financial Fraud from the IPO Prospectus

<p>In this study, I examine whether it is possible to predict future financial statement fraud using disclosure content prior to the fraud. Specifically, I employ a machine learning algorithm to construct a unique measure based on the lexical cues embedded within a firm’s first public disclosure, the Management’s Discussion and Analysis section of the S-1 filing, during the Initial Public Offering process. I use this measure to predict whether a firm that is not already committing fraud will commit fraud within five years of the Initial Public Offering (IPO) that results in an Accounting or Enforcement Release (AAER). I find there is information within the S-1 filing that is useful in the prediction of out-of-sample fraud. Additionally, I find that the measure performs better than both benchmark measures from prior literature and a new measure using quantitative information, when using information available at the S-1 date. Furthermore, the lexical cues measure performs well in predicting fraud relative to the benchmark measures even after updating the benchmark measures with misstated annual filings to aid their (but not my measure’s) fraud detection abilities. I find that my new measure is not limited to only predicting AAER based misconduct, but that the out-of-sample results hold when using an alternate sample based on 10(b)-5 filings as well as a comprehensive set of quantitative variables. Lastly, my measure identifies firms more likely to manage earnings to meet/beat analyst forecasts, firms who experience higher levels of information asymmetry around earnings announcements within the five years following the IPO, and has some predictive ability over future abnormal returns. </p>

Name thesis:degree_name

Doctor of Philosophy in Business Administration (PhD)

Level thesis:degree_level

Dissertation

Year dc:date.available

2021

Author dc:creator

• Anderson, Lisa Spadaccini

Advisor dc:contributor.advisor

• Allee, Kristian D.

Contributors dc:contributor

• Atwood, T.J.
• Richardson, Vernon J.

• Accounting or Enforcement Release (AAER)
• fraud_cues
• benchmark measures
• abnormal disclosure
• earnings management
• information asymmetry
• market returns
• fraud prediction
• Accounting
• Business Analytics
• Corporate Finance

Repository record dc:identifier

https://scholarworks.uark.edu/etd/4227

OAI identifier oai:identifier

oai:scholarworks.uark.edu:etd-5777