Nonlinear forecasting of euro area industrial production using evolutionary approaches
Title: Nonlinear forecasting of euro area industrial production using evolutionary approaches
Publisher: Springer Science+Business Media
Citation: Computational economics, 2018, Vol. 52, No. 2, pp. 521–530
ISSN: 0927-7099; 1572-9974
Stock Watson (in: Mills T, Patterson K (eds) Palgrave handbook of econometrics, Palgrave MacMillan, Basingstoke, 2003) argue that robust forecastability is dependent upon the optimality of the estimated parameters. Whilst recent studies in macroeconomic forecasting report the superiority of nonlinear models, yet they still suffer from precise parameter estimation. Our approach introduces evolutionary programming to optimize the parameters of various Threshold Autoregressive models. We generate forecasts for industrial production and compare our results versus linear benchmarks and quasi-maximum likelihood estimates for three Euro area countries. Based on our robust method, central banks and policy-makers could dynamically adjust their monetary and fiscal policy predictions.
Subject: Growth forecasting; Nonlinear models; Evolutionary methods; C32; C58; E2
Published online: 22 May 2017
Files in this item
There are no files associated with this item.