Composite Absolute Value and Sign Forecasts

48 Pages Posted:

See all articles by Andre B.M. Souza

Andre B.M. Souza

Universitat Pompeu Fabra - Department of Economics and Business; Barcelona Graduate School of Economics (Barcelona GSE)

Date Written: October 16, 2020

Abstract

This paper introduces composite absolute value and sign (CAVS) forecasts, a nonlinear framework that combines forecasts of the sign and absolute value of a time series into conditional mean forecasts. In contrast to linear models, the proposed framework allows different predictors to separately impact the sign and absolute value of the target series. Among other results, I show that the conditional mean can be accurately approximated by the product of mean squared error optimal sign and absolute value forecasts. An empirical application using the FRED-MD dataset shows that CAVS forecasts substantially outperform linear forecasts for series that exhibit persistent volatility dynamics, such as output and interest rates. The empirical application highlights that exploiting nonlinearities in macroeconomic series improves forecast accuracy.

Keywords: Forecasting, Directional Predictability, Machine Learning

JEL Classification: C53,C22,C52,C58

Suggested Citation

Souza, Andre B.M., Composite Absolute Value and Sign Forecasts (October 16, 2020). Available at SSRN: https://ssrn.com/abstract=

Andre B.M. Souza (Contact Author)

Universitat Pompeu Fabra - Department of Economics and Business ( email )

Barcelona
Spain

HOME PAGE: http://andrebmsouza.com

Barcelona Graduate School of Economics (Barcelona GSE) ( email )

Ramon Trias Fargas, 25-27
Barcelona, Barcelona 08005
Spain

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
3
Abstract Views
19
PlumX Metrics