This study uses a time-varying model that provides new evidence on the changing relationship between domestic overnight stays of selected winter sport destinations and natural snow conditions. A Kalman filter method combined with wavelet-based multiresolution analysis (MRA) is employed to investigate the relationships in intervals between 2 and 4 and up to 16-32 months. The model is applied to domestic overnight stays for selected mountain regions in Sweden (Dalarna and Jämtland), Norway (Buskerud, Hedmark, Hordaland and Oppland) and Austria (Salzburg and Tyrol). Results show that the sensitivity of domestic overnight stays on natural snow conditions varies markedly depending on location, time period and frequency band window used in the estimation. The medium-run relation for Tyrol and Salzburg is significantly declining over time, while in Norway and Sweden, the same relationship is generally volatile and not significant at the end of the sample period. In the short run, none of the regions exhibits a link between domestic overnight stays and snow depth fluctuations.
Keywords: Domestic tourism demand; Overnight stays; Snow depth; Time-varying models; Winter tourism.