Application of Box-Jenkins Models to the Tourist Inflow in Bhutan

Abstract

Bhutan has now increasingly become a popular destination for many international tourists. Tourism in Bhutan is considered as one of the largest foreign earning industries. The number of tourist inflow in the country is increasing year by year. Forecasting is very necessary for administration and tourist agent for creating awareness and planning for the future development. It can also predict the future trends as accurately as possible and helps in staying one step ahead of the competition. This study aims to apply mathematical model for forecasting monthly tourist inflow from Malaysia, Singapore, China, USA, England, France, Germany, Thailand, Australia and Japan to Bhutan. The Box-Jenkins model is used to identify the parameters of Autoregressive integrated moving average (ARIMA) model of monthly tourist visited data of above mentioned countries in the period 2011-2015 obtained from Tourism Council of Bhutan. An Akaike's Information Criterion, Schwartz's Bayesian Criterion and estimate variance of white noise are used throughout to test for the identification of best fit model. Further, the periodogram analysis was used to confirm the seasonal period of the model. The results showed ARIMA model for Thai, Chinese, Malaysian and Japanese, while seasonal ARIMA for American, Australian, British, French, Singaporean and German. Further, seasonal ARIMA model was obtained as the best fit model for the overall data. These models are illustrated and could possibly forecast the monthly tourist inflow of one year ahead with acceptable accuracy.