Mr. Jesus R Gonzalez-Garcia and Mr. Gonzalo C Pastor Campos
This paper examines the usefulness of testing the conformity of macroeconomic data with Benford's law as indicator of data quality. Most of the macroeconomic data series tested conform with Benford's law. However, questions emerge on the reliability of such tests as indicators of data quality once conformity with Benford's law is contrasted with the data quality ratings included in the data module of the Reports on the Observance of Standards and Codes (data ROSCs). Furthermore, the analysis shows that rejection of Benford's law may be unrelated to the quality of statistics, and instead may result from marked structural shifts in the data series. Hence, nonconformity with Benford's law should not be interpreted as a reliable indication of poor quality in macroeconomic data.
We introduce unFEAR, Unsupervised Feature Extraction Clustering, to identify economic crisis regimes. Given labeled crisis and non-crisis episodes and the corresponding features values, unFEAR uses unsupervised representation learning and a novel mode contrastive autoencoder to group episodes into time-invariant non-overlapping clusters, each of which could be identified with a different regime. The likelihood that a country may experience an econmic crisis could be set equal to its cluster crisis frequency. Moreover, unFEAR could serve as a first step towards developing cluster-specific crisis prediction models tailored to each crisis regime.