Alcohol Consumption in Australia: An Application of the Ordered Generalised Extreme Value Model

The adverse effects of excessive alcohol consumption are well-known. Of great concern to policy makers is to understand the potentially different drivers for consumers of different levels of alcohol consumption. Using unit record data from the Australian Drug Strategy Household Surveys, this paper estimates an Ordered Generalised Extreme Value model to identify the factors that influence differing levels of alcohol consumption. Unlike previous studies using inflexible approaches such as Ordered Probits/Logits or Multinomial Logits, the OGEV model is both flexible and consistent with random utility maximization. The results suggest that important drivers are: age; income; education; gender; and own and cross-pric