DEP (Socioeconomics) Discussion Papers, Macroeconomics and Finance Series 6/2014
Most macroeconomic data is continuously revised as additional information becomes available. We suggest that revisions of data is an important source of uncertainty about the state of the economy. This paper evaluates the quality of major real macroeconomic Euro area variables, published by Eurostat since 2001. The real time data set contains 159 vintages, covering the period of January 1991 until March 2014. The information content or informativeness of revision is measured using three methods: descriptive error statistics, signal-to-noise ratios and entropy measures. Our results document a trend of growing data uncertainty over the past decade for Euro area variables. As a robustness check, we reckon our results using US data and additionally show that uncertainty calculations are robust towards changes in final revision definition. Moreover, Euro area signal-noise-ratios and entropy measures are correlated with popular uncertainty proxies, Euro area news-based EPU and the VSTOXX. Our finding corresponds to the recent literature on increased macroeconomic uncertainty and especially economic policy uncertainty during and after the “Great Recession”.
forecasting information content uncertainty revisions revision errors entropy signal-to-noise ratio integrated signal-to-noise ratio recession EPU VSTOXX