Relationship between country risk volatility and indices based on unstructured information


  • Martín Llada Universidad de Buenos Aires


This work assesses whether certain indicators constructed from unstructured information published in newspapers contain useful information regarding dynamics of Argentina’s country risk volatility, estimated from a GARCH(1,1) model. The analysis covers the period 1998-2019. One standard deviation increment in the indicator that captures manifestations of pessimism is followed by an increment of approximately 0.2% in expected country risk volatility in the consecutive quarter. Out-of-sample exercises confirm that these non-traditional indicators allow for gains in forecast accuracy. These findings are robust to changes in the set of predictors, the specification of the model and the incorporation of new media content.


Macroeconomic forecasting, natural language processing, uncertainty, country risk volatility