Abstract: | Market operators monitor a massive flow of macroeconomic news every day and react to the unexpected component of each release. Can we replicate the market’s pricing of macroeconomic news automatically? This paper shows that a “Nowcasting Surprise Index”, constructed by aggregating forecast errors from a nowcasting model using model-based weights, resembles the surprise indexes proposed in the recent literature or constructed by practitioners, which cumulate survey-based forecast errors weighted by using the average effect of news on asset prices. This suggests that market operators and a nowcasting model filter the macroeconomic data flow similarly and confirms the link between news about macroeconomic indicators and asset prices. Moreover, the paper shows that recent cumulated news in macroeconomic data, which carry information about the underlying state of the economy, accounts for a non-negligible part of asset price behaviour. |