Nedenfor har jeg vedhæftet mine bacheloropgave på økonomi uddannelsen på KU. I opgaven undersøger jeg, hvorvidt Google søgninger fortsat prædikterer afkast på det amerikanske aktiemarked.
Abstract:
“We revisit the predictive relationship between Google-search-based investor sentiment and short-run asset prices and volatility in a market environment with increased retail participation and online investment communities. Using data from 2020-2024, we develop X-FEARS, a modernized version of the Financial and Economic Attitudes Revealed by Search (FEARS) index, combining traditional economic terms with contemporary retail-investor language such as buy the dip, and show that the classic pattern of sentiment-driven overreaction followed by short-run reversal is not a stable result. In 2020-2021, fear-related searches predict next-day reversal, in 2022 the effect disappears, and in 2023-2024 they predict continued market decline. Neither reversal nor continuation is concentrated among hard-to-arbitrage stocks. Further, X-FEARS has a positive contemporaneous effect on implied volatility, while its short-run dynamics qualitatively mirror the return results: reversal in 2020-2021, no effect in 2022, and continuation in 2023-2024. Finally, the co-movement between sentiment volatility and return volatility remains strong throughout, suggesting that sentiment and market instability continue to move together even as the short-run correction mechanism has weakened. Overall, our results show that search-based sentiment continues to predict asset prices and volatility, but that the short-run correction mechanism depends on the market environment in which sentiment is expressed.”
