TY - GEN
T1 - Firm risk identification through topic analysis of textual financial disclosures
AU - Zhu, Xiaodi
AU - Yang, Steve Y.
AU - Moazeni, Somayeh
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2017/2/9
Y1 - 2017/2/9
N2 - Corporate risk disclosures as part of U.S. public companies' financial reports are mandated by the Securities and Exchange Commission (SEC) since 2005. It provides forward-looking information about companies' future business and potential risks. This study analyzes risk types revealed in these risk disclosures and examines their potential implications on stock returns. Using 16,110 risk disclosures submitted to the SEC from 2011 to 2015, we apply Sentence Latent Dirichlet Allocation (Sent-LDA) model to infer risk types and propose a novel algorithm to match new factors with existing risk types which generates 90% correct matches. We then quantify the impact of different risk factors on the distribution of stock returns using different time windows. We find that common risk factors, such as accounting risk and acquisition risk, have significant effects on both long-term and short-term stock returns. Some other factors only have short-term or long-term effects on stock returns. These findings provide evidence that the companies' self-disclosed risk factors have significant impacts on subsequent stock return volatility and such impacts can be used to predict potential stock change after the public release of the financial risk disclosures.
AB - Corporate risk disclosures as part of U.S. public companies' financial reports are mandated by the Securities and Exchange Commission (SEC) since 2005. It provides forward-looking information about companies' future business and potential risks. This study analyzes risk types revealed in these risk disclosures and examines their potential implications on stock returns. Using 16,110 risk disclosures submitted to the SEC from 2011 to 2015, we apply Sentence Latent Dirichlet Allocation (Sent-LDA) model to infer risk types and propose a novel algorithm to match new factors with existing risk types which generates 90% correct matches. We then quantify the impact of different risk factors on the distribution of stock returns using different time windows. We find that common risk factors, such as accounting risk and acquisition risk, have significant effects on both long-term and short-term stock returns. Some other factors only have short-term or long-term effects on stock returns. These findings provide evidence that the companies' self-disclosed risk factors have significant impacts on subsequent stock return volatility and such impacts can be used to predict potential stock change after the public release of the financial risk disclosures.
UR - http://www.scopus.com/inward/record.url?scp=85016078068&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85016078068&partnerID=8YFLogxK
U2 - 10.1109/SSCI.2016.7850005
DO - 10.1109/SSCI.2016.7850005
M3 - Conference contribution
AN - SCOPUS:85016078068
T3 - 2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016
BT - 2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016
T2 - 2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016
Y2 - 6 December 2016 through 9 December 2016
ER -