A Dynamic Correlation Approach of the Swiss Tourism Income

The tourism sector is one of the most significant sectors in the modern world economy. However, despite its significance, the economics of tourism has not been given much attention, at least when compared with more core economics areas such as macroeconomics or econometric theory and methods, (Papatheodorou 1999). Furthermore, within the economics of tourism literature, econometric tools are rather limited, for example, in comparison to those applied in macroeconomics. However, in recent years, the number of papers using econometric methods and tools in tourism research has increased significantly. Several authors already employ standard econometric tools such as ARIMA modeling, Cointegration and Error Correction Mechanisms for forecasting purposes and to measure the long-run relationship between tourism and GDP, and when data is not available, or of low quality, Computable General Equilibrium models are implemented to assess the impact of tourism on other sectors. See, inter alia, Ballaguer and Catavella-Jorda (2002), Dritsakis (2004), Durbarry (2004), Papatheodorou and Song (2005), Narayan (2004), Sugiyarto et al. (2003), Wyer et al. (2003). Reviewing the relevant literature one can realize that the vast majority of econometric research in tourism is conducted almost exclusively in the time domain while frequency-domain (spectral and cross-spectral) methods are rather the exception. For example, out of 121 studies referring to modeling and forecasting of the tourism demand, only one (Coshall 2000) apart from seasonality modeling, applied frequency-domain analysis, as it is evident from a review made by Song and Li (2008) of post 2,000 research papers on the issue. In his research, Coshall (2000) found that cycles of passenger flows from UK to France, Belgium and The Netherlands depend on cycles in exchange rates, not on the GDP cycle.