Title: The switching skewness over the business cycle
Authors: Stephane Lhuissier - Banque de France (France) [presenting]
Abstract: Motivated by the analysis of the evolution of the distribution of macroeconomic time series data over time, the aim is to develop and apply a Gibbs-sampler for autoregressive time series subject to regime switches in the tails of the distribution. More specifically, we consider the skew-normal distribution, in which the shape parameter is allowed to change over time according to a Markov-switching process. As an empirical illustration, we analyse the distribution of the growth rates of postwar U.S. real GDP, and find periodic shifts between a left and right-skewed distribution regime, with the former corresponding closely to NBER recession dates. Hence, more theorizing is needed to better understand the interaction between variation in tails and the business cycle.