DP15109 Modeling and Forecasting Macroeconomic Downside Risk
|Author(s):||Andrea De Polis, Davide Delle Monache, Ivan Petrella|
|Publication Date:||July 2020|
|Keyword(s):||Business cycle, Downside risk, financial conditions, score driven models, Skewness|
|JEL(s):||C53, E32, E44|
|Programme Areas:||Monetary Economics and Fluctuations|
|Link to this Page:||cepr.org/active/publications/discussion_papers/dp.php?dpno=15109|
We investigate the relation between downside risk to the economy and the financial markets within a fully parametric model. We characterize the complete predictive distribution of GDP growth employing a Skew-t distribution with time-varying location, scale, and shape, for which we model both secular trends and cyclical changes. Episodes of downside risk are characterized by increasing negative asymmetry, which emerges as a clear feature of the data. Negatively skewed predictive distributions arise ahead and during recessions, and tend to be anticipated by tightening of financial conditions. Indicators of excess leverage and household credit outstanding are found to be significant drivers of downside risk. Moreover, the Great Recession marks a significant shift in the unconditional distribution of GDP growth, which has featured a distinct negative skewness since then. The model delivers competitive out-of-sample (point and density) forecasts, improving upon standard benchmarks, especially due to financial conditions providing a strong signal of increasing downside risk.