Econometric Models of Alcohol Demand in Danish Households
Jesper Riis-Vestergaard Sorensen
University of California, Los Angeles (UCLA) - Department of Economics
May 4, 2011
Limited dependent variables are common in cross-section analysis of alcohol demand. The standard tool, the censored regression or tobit model, implies an undesirable restriction on household behaviour, which can be relaxed by using the double-hurdle model. The double-hurdle model is extended by incorporating the Box-Cox transformation, nesting a range of alternative models. Household expenditures on alcohol are analyzed with the use of Statistics Denmark’s 2005 Household Budget Survey. Estimation of budget elasticities indicates that the tobit specification leads to unreliable results. Results from the preferred Box-Cox double-hurdle are in line with the literature.
Number of Pages in PDF File: 18
JEL Classification: C30, C31, C34, D12
Date posted: November 20, 2011