Forecasting Foreign Trade of Bosnia and Herzegovına for Wood and Articles of Wood, Wood Charcoal by Seasonal ARIMA Model

Nadi̇r Ersen,İlker Akyüz

doi:10.21533/pen.v5i1.73

Abstract

In this study, it is aimed that the analysis of export and import values of Bosnia and Herzegovina for wood and articles of wood, wood charcoal with seasonal ARIMA model and forecasting of export and import values for next term by the best appropriate seasonal ARIMA model. The data used in this study were obtained from Trade statistics for international business development (TRADEMAP) and monthly data covering the period of January 2007 and December 2015. Augmented Dickey-Fuller test was used for the stationarity test. Temporary model that have smallest values of forecasting accuracy measurement was determined. The appropriateness of the model (whether plot of autocorrelation has white noise) was determined by using the Box-Ljung test. As a result, ARIMA(3,1,0)(0,1,2)12 model was found as the best forecasting model for both export and import series. It was estimated that export value of Bosnia and Herzegovina for wood and articles of wood, wood charcoal is approximately 531 million$, while import value is 160 million$ in 2020. Key words: Seasonal ARIMA model, wood and articles of wood, wood charcoal, export, import, forecasting

Highlights

Export provides to remain in the balance of foreign trade by supply the amount of foreign currency going abroad
It was seen that series is stationary in the 5% significance level according to Augmented Dickey-Fuller (ADF) test when taken difference of the natural logarithm series
The export and import values of Bosnia and Herzegovina for wood and articles of wood, wood charcoal with seasonal ARIMA model were estimated with seasonal ARIMA model

Summary

Introduction

Export provides to remain in the balance of foreign trade by supply the amount of foreign currency going abroad. The forecasting of accurately value of export/import is very important. Quantitative methods are contained in the method of time series analysis [1]. Box-Jenkins method is one of the most used methods This method is called as ARIMA (autoregressive integrated moving average) analysis. ARIMA model combine differencing with autoregression (AR) and a moving average (MA) models. This model can be expressed as wt=c+φ1wt-1+......+φpwt-p +θ1εt-1 +... Wt is the differenced series, p is order of the autoregressive part and q is order of moving average part. We said this an ARIMA (p,d,q), where, d is a degree of difference [2].

Methods

Results

Conclusion