Prediction of international tourist arrivals to Sri Lanka using state space modelling method: An empirical study after the civil war

Abstract

Tourism plays a big role in development of a country in terms of economics. It contributes in large scale to the national revenue in Sri Lanka through foreign exchange. Tourists had less interest of visiting Sri Lanka, mainly due to the uncertainty of security during the civil war. Previous studies identified some models to forecast the arrivals to Sri Lanka but none of them have used state space modelling method. In Sri Lanka the war is over now and the security is certain in all over the island. Thus the tourist arrivals are getting increase in recent past. Therefore, the objective of the study is to fit the model to forecast the international tourist arrivals, after the civil war, to Sri Lanka by using state space modelling method. Prior to model fittings, preliminary testing is carried out to study the behaviour of the series. Augmented Dickey- Fuller (ADF) and Kruskal- Wallis tests are applied to confirm the stationary of the series. In state space modelling, Kalman filter technique is used to estimate the parameters of the model. Diagnostic checking for the fitted model is also carried out. To check the accuracy of fitted model mean absolute percentage error (MAPE) statistics is employed. After the civil war a dramatic increase is clearly observed with positive growth rates every year. Until March 2016, the total arrivals are over 7.961 million. The MAPE statistics of the fitted state space model is 5.95%. By ex-post forecast, the estimated arrivals for the next 3 quarters of the year 2016 is approximately 1.613 million and which is nearly 22.2% increase in growth with the year 2015. Therefore, the fitted model can be used to predict the future arrivals to Sri Lanka and accordingly the policy maker can take necessary action to cater them in future.