A Bivariate Integer Valued Allocation Model for Guest Nights In Hotels and Cottages

Abstract

The number of Norwegian guest nights in Swedish hotels and cottages is studied. Aggregation of an integer-valued AR(1) model and a two-stage demand model underlies the empirical results. The parameters in the model are check-out probability, mean check-in and the probability of selecting the hotel alternative. These parameters are specified to depend on economic variables implied by demand analysis. Via the probability of selecting a hotel, the empirical results indicate a substitution towards less expensive accommodation as the Swedish price level increases. For the check-out probability, an increase in the cottage price reduces the probability for staying another night in cottage, whereas an increase in the hotel price indicates a decrease in the check-out probability for hotel.

Keywords: Integer-valued time series, demand analysis, tourism, accomodation, hotel, cottage

JEL Classification: C32, C35, C51, D12