Baker, Bloom, and Davis, (2016) “Measuring Economic Policy Uncertainty” created their Economic Policy Uncertainty (EPU) indices for various countries by counting articles about uncertainty, the economy, and policy. Their methodology was to classify the articles by a set of pre-determined keywords. The following two research papers create similar uncertainty indexes but differ in the way they classify news articles.
Vegard Høghaug Larsen in this papers shows that depending on the source, uncertainty may have different effects on the same macroeconomic variables. In order to create his uncertainty measure he uses machine learning techniques to classify news articles published in daily business newspapers. This method allows him to identify different sources of uncertainty and evaluate their effect on the economy.
Components of Uncertainty
Author: Vegard Høghaug Larsen
From: Norges Bank
Calomiris and Mamaysky follow a different news classification method: they use an atheoretical method to organize the flow of words in news articles. They find that their word flow measures namely sentiment, frequency, and entropy can predict future country-level returns, volatilities, and drawdowns.
How News and Its Context Drive Risk and Returns Around the World
Authors: Charles W. Calomiris and Harry Mamaysky
From: Columbia University
Song and Tang present a two-state Markov-switching growth model where news helps agents to learn the true values of parameters. They find that news can actually increase uncertainty about future growth when it contradicts agents’ existing beliefs. In particular bad news tends to persist during the recovery phases following recessions.
News-Driven Uncertainty Fluctuations
Authors: Dongho Song, Jenny Tang
From:Boston College – Federal Reserve Bank of Boston
