Determinants of tourism hotel expenditure in Hong Kong

This paper reports on a study about applying regression techniques to identify the determinants and functional forms of tourism hotel expenditure in Hong Kong. Annual time series data from 1983 to 1997 of average room rate, the number of visitor arrivals, the service price index, and hotel accommodation rates were hypothesised to affect tourism hotel expenditure. Seven exogenous variables were selected for regression model development in both linear and log‐linear forms. In view of the potential problems of multicollinearity between the independent variables, and therefore the associated instability of the regression coefficients, stepwise regression analyses were employed to improve the initial model. Final empirical results showed that the hotel expenditure in Hong Kong could be explained by four of the seven exogenous variables. A log‐linear form of the regression model appeared to slightly outperform the linear form.