The Determinants and Stock Market Consequences of Forward-Looking Statements: A Machine Learning Textual Analysis Approach

Student thesis: Doctoral ThesisDoctor of Philosophy


This thesis aims to analyse the proprietary costs and the potential market consequences of the firm’s voluntary disclosures. I use a particular disclosure of corporate filings, i.e., the Forward-Looking Statements (FLSs) from corporate 10-K or 8-K filings. The conflict exists between the SEC’s appeal for more informative disclosures and the managers’ incentives of withholding. While proprietary costs can rationally explain the manager’s decision in face of such conflict, whether the disclosure of FLSs is affected by such costs is largely uncertain.

Another uncertainty in disclosure theories is regarding the market consequences of the FLSs disclosure. The stock price synchronicity is believed to measure the firm-specific (market-wide) information flow to investors. Whether the disclosure of FLSs affects investors’ perception of firm-specific information is still an open question. Even if the disclosure of FLSs does affect investors’ perception of firm-specific information flow, it is unknown whether the manager strategically uses FLSs to manipulate investors’ perception.

I first investigate the proprietary costs related to FLSs using the states' staggered adoption of the Inevitable Disclosure Doctrine (IDD). Specifically, whether FLSs contain proprietary information and which type of proprietary information does FLSs contain are somewhat equivocal. Based on the sample of 24,094 U.S. firms from 1994 to 2010, our difference-in-differences (DID) test results show that managers reduce the FLSs in MD&A (especially the non-earnings and qualitative FLSs) in response to states' adoption of IDD. We also find that the tone of FLSs is more favourable for firms in states adopting IDD than other firms. Such finding is consistent with our hypothesis that the risk of losing firm's trade secrets is reduced by the states' adoption of IDD, so managers are less likely to discuss the potential risks of losing trade secrets (by Regulation S-K).

Then, this thesis analyses the market effects of FLSs in 8-K filings on stock price synchronicity. Based on a sample of US firms between 2004 and 2020, I find that both the intensity and the tone of FLSs are negatively related to future stock price synchronicity, implying that more (favourable) FLSs increase the firm's accounting transparency, which helps investors perceive firm-specific information. Our findings highlight the role of FLSs in ameliorating firm's information environment, potentially via suppressing managers' incentive to conduct earnings management. Such finding provides new evidence that the FLSs can influence the investors' perception of the firm-specific information flows.

In order to find whether managers utilize FLSs to conduct strategic reporting, I further test the disclosure of external attributions in FLSs of 8-K filings from 2004 to 2020. I find that the higher proportion of external attributions increases the managers’ bad news hoarding, which increases stock price crash risks. There is also a positive joint effect of disclosing external factors and earnings management on stock price crash risks. Taken together, this thesis provides implications for regulators and economic agents.

For regulators such as the SEC committee or the US states’ court, this thesis potentially offers insights that the disclosure of FLSs is affected by proprietary costs, and encouraging corporate to disclosure of FLSs may impair firm’s competitive advantage but benefit investors’ perception. For economic agents, this thesis highlights that the disclosure of FLSs can be strategically used by managers, depending on the type of the FLSs (e.g., forecast-like vs others, or external factors). Therefore, investors ought to carefully identify the disclosure type of FLSs and exercise judgements accordingly.
Date of Award1 Jul 2023
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorMingzhu Wang (Supervisor)

Cite this