Modeling and forecasting Hong Kong tourist arrivals
Student thesis: Master's Thesis
Related Research Unit(s)
In recent years, the government of the Hong Kong SAR intends to diversify the economic structure and make it a more knowledge-based and high value-added economy. Tourism is of course one of the key industries that the government focuses to develop. To grasp the growth of the international tourist arrivals, better infrastructure and wellorganized marketing strategies are necessary. Indeed, it is essential to forecasting Hong Kong tourism demand so as to generate an efficient planning and policy making for accommodating the possible changes in tourist arrival pattern. This research aims to forecast the monthly tourist arrival in Hong Kong and predict the effect on tourist arrival pattern due to the opening of Hong Kong Disneyland. Statistical techniques in time series analysis, namely naïve, moving average, exponential smoothing, autoregressive integrated moving average approach, vector autoregressive model and time-varying parameter analysis, are constructed to understand the arrival pattern and forecast the tourism demand. Explanatory variables (exchange rate and relative price level) are employed to investigate their relationship with Hong Kong tourism demand. Unit root tests (augmented Dickey Fuller test and Hylleberg, Engel, Granger and Yoo test) are employed to investigate the presence of ordinary and seasonal unit root. Historical monthly tourist arrival data from January 1981 to December 2003 are collected to build models and two accuracy measures (mean absolute percentage error and root mean squared error) are adopted to evaluate forecasting performance of various methods. The result indicates that simple forecasting methods like naïve and decomposition models outperformed the more sophisticated approaches like transfer function model.
- China, Tourism, Visitors, Foreign, Forecasting, Hong Kong