You are here
The expected time to cross a threshold and its determinants: A simple and flexible framework
C32 - Time-Series Models
C41 - Duration Analysis
C51 - Model Construction and Estimation
In this paper we introduce a flexible framework to estimate the expected time (ET) an outcome variable takes to cross a threshold conditional on covariates. The proposed methodology makes use of the Markovian property and allows us to infer the impacts that covariates have on the ET an outcome variable takes to revert to a value of interest (for instance, its mean) given a specific starting point. An empirical application to the U.S. economy is provided, which investigates how the yield spread (YS) influences the ET the industrial production (IP) growth rate takes to return to its mean considering several initial values for the outcome variable. Our results suggest that the YS may have an important role in stimulating a faster return to desirable growth rates when the economy is in contraction or faces weak growth. Moreover, the YS also seems relevant in the presence of positive and high IP growth rates since a negative value of this variable may contribute for the IP growth rate to quickly return to below average values.