Cruise ship passenger spending patterns are analyzed for the archipelago of the Canary Islands, as such patterns represent a key element in the evaluation of the economic impact of cruise ships on residents and local stakeholders. Over six cruise seasons, data regarding cruise passenger expenditures were collected via survey responses at each stopover. Since the five categories of expenditure analyzed are censored and possibly correlated, we have estimated a multivariate tobit system. This approach offers more efficient estimates of the determinants of cruise passengers' onshore spending, which can be useful in designing economic policies. Our results show that gender, age and socioeconomic status affect cruise passenger spending patterns significantly and, more importantly, in different ways depending on the expenditure category. Therefore, it seems that more personalized marketing strategies (i.e., gender/age/nationality-oriented) classified by expenditure category should be more efficient and, therefore, implemented to achieve greater local economic impact.