An Econometric Forecasting Model for Describing Tourist Arrivals Process in Sri Lanka

Authors

  • Hemantha P. Diunugala Department of Social Statistics, University of Sri Jayewardenepura

Abstract

Tourism is one of the major potential growth sectors in Sri Lanka. It contributesimmensely to the economic growth of the country. This contribution has been quitesignificant during the past four decades. Although the tourist arrivals increased rapidlyafter introducing an open economic policy in 1977, the growth of tourism in Sri Lankahas not been steady and smooth. The study is based on the tourist arrivals‟ data from thehighest tourist arrivals countries to Sri Lanka such as India, United Kingdom (UK),Germany, Maldives, France, Australia, Netherlands, United States of America (US),Japan and Italy during the period from January 1977 to April 2012. The tourist arrivalsseries of each country exhibits different cyclical periods and trend patterns, strongseasonality, and irregular fluctuations. The aim of the study is to find the best fitted timeseries model for describing tourist arrivals process to Sri Lanka from each country.

The Winter‟s Multiplicative Exponential Smoothing Method (WMESM) and Boxand Jenkins Multiplicative Seasonal Auto Regressive Integrated Moving Average(SARIMA) method were applied to describe tourist arrival processes. Standard modelselection criteria were used to select the best fitted models.

Most appropriate model for describing tourist arrivals processes of countries,India, UK, Germany and Australia are the WMESM on levels data while Maldives,Netherlands, USA, Japan and Italy are the WMESM on logarithmic transformed data.The SARIMA model on logarithmic transformed data is the most appropriate model onlyfor describing tourist arrival processes of France and it was ARIMA (1,1,1) (0,1,1)12.According to the scale developed by Lewis based on Mean Absolute Percentage Error(MAPE) models for Germany, Maldives Australia and USA (MAPE is less than 10%)are highly accurate whilst models for India, UK, France, Netherlands, Japan and Italy(MAPE is 10-20%) remain in a good level of accuracy.

Key words: Tourist arrivals, Best fitted model, Model selection criteria, Levels ofaccuracy

Author Biography

Hemantha P. Diunugala, Department of Social Statistics, University of Sri Jayewardenepura

Department of Social Statistics, University of Sri Jayewardenepura

Downloads

Published

2015-07-03