News perception in financial markets: the role of signaling, cognitive biases and heuristics in understanding financial disclosures

Abstract: Stock splits and their controversies are extensively studied. However, while many studies document post-split signaling evidence, little is known about the nature and form of management signals ahead of stock splits. Inspired by signaling theory, we extend previous research and unravel the role of language in Form 10-K and 10-Q performance reports as positivity signal from management ahead of stock splits. We create a multi-period signaling model and use sentiment analysis to assess language signals in management’s narrative on their firm’s current state and future performance outlook. Our results allude to a pre-split positivity signal. Splitting firms significantly increase the use of positive language over the year leading up to a split announcement (>10% more positivity). Our results further constitute positive signal confirmation after the split before a reversal of tone. This uptake in positive language is highly significant in explaining firm’s splitting likelihood and helps investors to anticipate splits in coming quarters and thereby future corporate decision-making. By and large, the market overreacts to the Form 10-K and 10-Q disclosures that are related to the pre-split positivity signal in the short-term and this overreaction is more pronounced after negative news. It is an intriguing notion to further study the affective characteristics of language signals in corporate communication and their interplay with corporate financial decision-making.

Noise trader models in financial markets distinguish two sets of trading activity based on rationality: (a) informed trading based on rational decisions and fundamental information and (b) noise trading based on non-fundamental noisy signals that lack a deeper meaning. Financial disclosures serve as key intermediaries between companies and the stock market, yet little is known about semantic drivers of information perception, particularly not in relation to these fundamental and non-fundamental decision-making patterns. This work extends previous research by unraveling the role of word choice, semantic orientation and hidden topic structures in U.S. regulated Form 8-K filings in relation to the two groups of trading activity. For this purpose, we use Bayesian filtering and supervised Machine Learning approaches of sentiment analysis to investigate disentangled effects of information perception in financial markets. We identify that news perception linked to non-fundamental decision-making is based on not discerning the full information breadth from financial disclosures and disparate interpretations of textual semantics as well as overall documents. We identify greater information perception ambiguity in relation to price noise, particularly for the case of fact- and emotion-laden content. The news perception differences are a cause for market prices to stray apart from fundamental values.

Scanning heuristics help individuals to optimize information processing effort in face of extensive amounts of new information. The deceptive character of news headlines, however, affects individual news perception and may stray from uncovering the true content of news articles. In absence of an underlying asset, the value and price formation of Bitcoin is majorly driven by investors’ beliefs and perceptions about new information and their derived expectations about the cryptocurrency’s future value. This work extends previous research by unraveling investors’ news perception of Bitcoin-themed online news and the implications of headline scanning heuristics on Bitcoin trading measures. Investors’ anchoring to news headlines explains short-term Bitcoin price reactions. Correspondingly, articles that convey opposing sentiment as their headline introduce uncertainty and are significantly linked to future Bitcoin price volatility

Angaben zum Objekt

• Standort: Deutsche Nationalbibliothek Frankfurt am Main
• Umfang: Online-Ressource
• Sprache: Englisch
• Anmerkungen: Universität Freiburg, Dissertation, 2019

Klassifikation und Themen

• Klassifikation: Wirtschaft
• Schlagwort: Markets ; Perception ; Comprehension

Beteiligte, Orts- und Zeitangaben

• Ereignis: Veröffentlichung
• (wo): Freiburg
• (wer): Universität
• (wann): 2020
• Urheber: Rupalla, Felix
• Beteiligte Personen und Organisationen: Neumann, Dirk ; Albert-Ludwigs-Universität Freiburg. Wirtschafts- und Verhaltenswissenschaftliche Fakultät

Weitere Informationen

• DOI: 10.6094/UNIFR/166862
• URN: urn:nbn:de:bsz:25-freidok-1668629
• Rechteinformation: Der Zugriff auf das Objekt ist unbeschränkt möglich.
• Letzte Aktualisierung: 25.03.2025, 13:56 MEZ