演講主題:Semantics Matter: An Empirical Study on Economic Policy Uncertainty Index
講題摘要:
When dealing with textual data, previous studies mainly use the keyword-based matching method to construct indices. The EPU index proposed by Baker et al. (2016) is one example. However, in this paper, we argue that due to its neglect of semantics, such keyword matching generates excessive noise, which affects index quality and further leads to incorrect inferences. We address this shortcoming by adopting the EPU index as an example. We investigate several neural network models and select the best-performing classifier to remove the noise caused by keyword matching. Our empirical results show that the de-noised EPU index is useful in predicting economic variables and generating superior out-of-sample forecasts. Furthermore, the effects of policy uncertainty shocks on core macro variables of interest are consistent with the predictions of macroeconomic theory. Because the proposed approach is a general framework, in the future all keyword matching-based indexes can be improved under the same approach.