Forecasting Macroeconomic Variables Using Artificial Neural Network and Traditional Smoothing Techniques

Journal of Applied Finance and Banking, Vol.3, No.4, 2013, pp.73-104.

26 Pages Posted: 14 May 2013 Last revised: 15 Jun 2013

See all articles by Emrah Önder

Emrah Önder

Istanbul University - School of Business - Department of Quantitative Techniques

Fırat Bayır

Istanbul University - School of Business

Ali Hepsen

Istanbul University - Faculty of Business Administration, Department of Finance

Date Written: May 13, 2013

Abstract

For many years, economists have been using statistical tools to estimate parameters of macroeconomic models. Forecasting plays a major role in macroeconomic planning and it is an essential analytical tool in countries’ economic strategies. In recent years, researchers are developing new techniques for estimation. Most of these alternative approaches have their origins in the computational intelligence. They have the ability to approximate nonlinear functions, parameters are updated adaptively. In particular, this research focuses on the application of neural networks in modeling and estimation of macroeconomic parameters. Neural networks have received an increasing amount of attention among macroeconomic forecasters because of the ability to approximate any linear and nonlinear relationship with a reasonable degree of accuracy. Turkey is one of the European Union candidate countries such as as Iceland, Montenegro, Serbia, The former Yugoslav Republic of Macedonia. In this study eight macroeconomic indicators including gross domestic product (volume, NGDPD), gross national savings (NGSD-NGDP), inflation (average consumer prices, PCPI), population (LP), total investment (NID-NGDP), unemployment rate (LUR), volume of exports of goods and services (TX-RPCH), volume of imports of goods and services (TM-RPCH) were used for forecasting. As analysis tools, classical time series forecasting methods such as moving averages, exponential smoothing, Brown's single parameter linear exponential smoothing, Brown’s second-order exponential smoothing, Holt's two parameter linear exponential smoothing and decomposition methods applied to macroeconomic data. The study focuses mainly on the applicability of artificial neural network model for forecasting macroeconomic parameters in long term and comparing the artificial neural network’s results with the Traditional Time Series Analysis (Smoothing & Decomposition Techniques). To facilitate the presentation, an empirical example is developed to forecast Turkey’s eight important macroeconomic parameters. Time Series statistical theory and methods are used to select an adequate technique, based on residual analysis. Turkey will celebrate the 100th anniversary of its foundation in 2023. Policies and implementations targeted for raising economic position.

Suggested Citation

Önder, Emrah and Bayır, Fırat and Hepsen, Ali, Forecasting Macroeconomic Variables Using Artificial Neural Network and Traditional Smoothing Techniques (May 13, 2013). Journal of Applied Finance and Banking, Vol.3, No.4, 2013, pp.73-104., Available at SSRN: https://ssrn.com/abstract=2264379 or http://dx.doi.org/10.2139/ssrn.2264379

Emrah Önder (Contact Author)

Istanbul University - School of Business - Department of Quantitative Techniques ( email )

Istanbul
Turkey

Fırat Bayır

Istanbul University - School of Business ( email )

Istanbul Business School
Avcilar Campus
Istanbul, 34322
Turkey

Ali Hepsen

Istanbul University - Faculty of Business Administration, Department of Finance ( email )

Istanbul
Turkey

Do you want regular updates from SSRN on Twitter?

Paper statistics

Downloads
319
Abstract Views
1,360
rank
130,720
PlumX Metrics