CMStatistics 2017: Start Registration
View Submission - CMStatistics
Title: Predicting recessions in Italy: A nonparametric discrete choice model for time series Authors:  Camilla Mastromarco - University of Salento - Lecce (Italy) [presenting]
Leopold Simar - Universite Catholique de Louvain (Belgium)
Valentin Zelenyuk - University of Queensland (Australia)
Abstract: Efficiency frontier is estimated for Italian economy using quarterly data from 1995 to 2016. A flexible nonparametric two-step approach on conditional efficiencies allows us to eliminate the dependence of production inputs/outputs on time. The efficiency measure can be interpreted as output gap and employed as a predictor of economic slowdown. Applying a generalised non-parametric quasi-likelihood method in the context of discrete choice models for time series data, we investigate how the spread variable, constructed as the difference between the 10-year German Treasury bond and 10-year Italian Treasury bond and our estimated efficiency scores predict the recession in Italy. By using a dataset from 1995 -2014 with quarterly frequency we emphasize the usefulness of this model for the prediction of Italian recessions in case of two explanatory variable (the lagged spread and efficiency scores). Our model involves two continuous predictors, the spread and the efficiency scores and one discrete variable, the lagged dependent variable. We find that this flexible nonparametric approach offers additional insights than the usual linear probit frequently used in the literature in this context.