Fake News and Market Volatility: Insights from a Large Language Model

Abstract

During the last decade, the spread of fake news using the internet has reached dramatic levels and with unknown consequences to individuals (Lazer et al., 2018). This paper explores (i) to what extent fake and real news contribute to the spread of financial market volatility and (ii) if fake and real news induce market volatility through comparable connectedness patterns.

Using the latest developments in machine learning to detect fake news, we empirically find that investors distinguish between what is fake and what is real. Real news has an enduring and substantial effect on investment decisions while fake news is transient and less significant. We conclude that fake news contributes to spreading market volatility, but that financial markets quickly and efficiently sort out fake from real. We contribute to a growing strand of literature examining the consequences of the recent increase in the spread of fake news through online platforms.