Twitter-based economic policy uncertainty index for Chile

Abstract

In this paper, we develop a daily-frequency measure of economic and policy uncertainty for Chile, employing information obtained from Twitter accounts using web scraping techniques and following closely the methodology proposed by Baker et al. (2016). Our proposed measure, called DEPUC, aims to capture the level of general disagreement –a proxy for economic and policy uncertainty– in topics such as the economy, economic policies, uncertainty about particular events, and Chile’s conjuncture situation. The index, available from 2012 onwards, shows significant hikes that coincide with several local and international episodes that provoked extraordinary levels of uncertainty in Chile, especially after the events around the civil protests in mid-October 2019 and the start of the COVID-19 pandemic in mid-March 2020. An empirical exercise reveals that the proposed measure is a significant determinant of the nominal exchange rate dynamics, especially when this variable’s magnitude is high and a week after the shock occurs.On the contrary, when the exchange rate is low, the impact of uncertainty on this variable is quantitatively smaller for any forecasting horizon. These features, and others discussed in the paper, highlight the usefulness of the proposed metric as an additional indicator that policymakers can incorporate into their monitoring toolkit.