A New Index for Measuring Uncertainty Due to the COVID-19 Pandemic

Abstract

This study contributes to the emerging literature offering alternative measures of uncertainty due to the COVID-19 pandemic. We combine both news-and macro-based trends to construct an index. The former involves the use of Google trends with plausible variants of words used to capture the pandemic, which are combined using principal components analysis to develop a news-based index. For the macro-based index, we identify global factors such as oil price, stock price, Dollar index, commodity index and gold price, and thereafter we obtain the macro-based uncertainty using variants of stochastic volatility models estimated with Bayesian techniques and using a dynamic factor model. Consequently, the new (composite) index is constructed by combining the news- and macro-based indexes using principal components analysis. Our empirical applications of the index to the stock return predictability of the countries hit worst by the pandemic confirm the superiority of the composite index over the existing news-based index in both the in-sample and out-of-sample forecast horizons. Our results are also robust to forecast horizons and competing model choices.