Costs associated with exit or disposal activities: A topic modeling investigation of disclosure and market reaction

Summary The Securities and Exchange Commission (SEC) mandates disclosure of exit or disposal activity events in 8‐K filings. We use Latent Dirichlet Allocation (LDA), a topic modeling method from computational linguistics, to investigate the possibility that substantively different event types may b...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Intelligent systems in accounting, finance & management finance & management, 2023-10, Vol.30 (4), p.173-191
Hauptverfasser: Cullinan, Charles P., Holowczak, Richard, Louton, David, Saraoglu, Hakan
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Summary The Securities and Exchange Commission (SEC) mandates disclosure of exit or disposal activity events in 8‐K filings. We use Latent Dirichlet Allocation (LDA), a topic modeling method from computational linguistics, to investigate the possibility that substantively different event types may be subsumed under Item 2.05, the SEC category for costs associated with exit or disposal activities. Our analysis reveals that four distinct topics are reported under the Item 2.05 umbrella category: (1) restructuring, (2) disposal of a line of business, (3) plant closings, and (4) layoffs/workforce reductions. We then investigate various aspects of these 8‐K filings. We find that the market reacts most negatively to workforce reductions that are reported in the absence of a broader strategic initiative. Subsequent amendments to the Item 2.05 8‐K filings are significantly more likely for restructuring initiatives, and significantly less likely for layoffs. Asset impairment charges most frequently accompany line‐of‐business disposals and plant closings. Our results demonstrate that there are meaningful differences between the event types reported within Item 2.05 filings and that LDA provides a useful means of differentiating among these event types.
ISSN:1550-1949
2160-0074
DOI:10.1002/isaf.1545