Forecasting tourism demand in Chile: Regional analysis using the Seasonal Autoregressive Model

This paper presents Chilean tourism demand describing its behavior for both the country and each of its regions, the analyzed period comprises 2014:01 to 2019:02. The seasonal autoregressive model (SARIMA) process was used to model the series growing dynamics. Results show that best-fitting models capture nonlinear growth, seasonal patterns, and series volatility, and make it possible to describe not so obvious behaviors, such as the seasonal process order or long-term growth trends. From a public policy point of view, this provides relevant information for decision-makers to manage touristic services and infrastructure in a better way. Regional and countries’ forecasted demand presents a low error percentage, less than 2%, though in some regions this value is underestimated overestimated in others.