On the Impact of Financial, Newspaper, and Twitter-Based Uncertainty on U.S. Banks : A Comparative Multi-Scale Analysis

This study examines the multi-scale impact of financial, newspaper, and Twitter-based uncertainty on bank stock returns in the United States. By combining Principal Component Analysis and Discrete Wavelet Transformation, we estimate directional spillover and causality for different time horizons. The results reveal that although newspaper and Twitter-based measures are highly correlated, they differ in their impact on the banking sector. While Twitter affects the stock markets in the short run, economic policy uncertainty is important in the medium run (8-32 trading days). Moreover, short term feedback loops between uncertainty and the stock markets are found. Considering the impact size, VIX is the most influential factor. This demonstrates that the source and multi-scale consideration of uncertainty is important for assessing its impact