A Survey of Alternative Measures of Macroeconomic Uncertainty : Which Measures Forecast Real Variables and Explain Fluctuations in Asset Volatilities Better?

In the past twenty years, measures of economic uncertainty have been developed that are either purely market price-based, structural model-based using data on real fundamentals and asset prices, text-based, or survey-based. We compare the performance of these uncertainty measures in forecasting three real variables with irreversibilities: investment, hiring, and credit creation, as well as in explaining fluctuations in stock-market and Treasury bond market volatility. In general, we find that structural model-based measures do better than measures constructed using other approaches, with David and Veronesi (2013) model's stock market volatility performing the best on several (but not all) dimensions. Their learning-based model's volatility places time-varying weights on inflation, earnings, and consumption news, as agents in the economy assess the impact that inflation has on the stability of real economic growth. A measure of the uncertainty of a recession in the following quarter based on the Survey of Professional Forecasters' ``anxiety'' index has a fairly good performance, however, it is more volatile than the David and Veronesi (2013) stock market volatility