Forecasting Foreign Direct Investment Inflow to Sri Lanka: Hybrid ARIMA-Neural Network model
No Thumbnail Available
Date
2011
Journal Title
Journal ISSN
Volume Title
Publisher
Uva Wellassa University of Srilanka
Abstract
Note: See the PDF Version
FDI is known as that "Investment made by multinational business enterprises in foreign countries to control assets and manage production activities in those countries" (Bashier and Talal, 2007). Foreign direct investment consists of three major components. Those are equity capital, reinvestment earnings and other capital used in various intra company debt transactions. In 1977, the Sri Lankan government made changes in its economic policies to establish an investment friendly economic environment. Sri Lanka reached its highest FDI inflow of US dollars 752.2 million in 2008.
The objective of the research is to build a model to forecast future FDI inflow to Sri Lanka.
Research has been carried out using different methods such as the univariate ARIMA model (Bashier and Talal, 2007), the double exponential smoothing model (Kumar et al., 2009) and the neural network model (Pradhan, 2010) to forecast FDI inflow to other countries. To the best knowledge of the authors, this study could be the first to forecast FDI inflows in Sri Lanka using a combination of an ARIMA model and a neural network model. In this study, an ARIMA model is fitted through the Box-Jenkins procedure and then the back propagation neural network procedure is applied to remove the lack of accuracy due to the small sample size. The data set employed consists of annual foreign direct investment inflow to Sri Lanka from 1978 to 2010.
Description
Keywords
Marketing, Economics, Management, Business Studies