Enometrics XX - Household Consumption of Alcoholic Beverages: Estimation of Price and Expenditure Elasticities using a Demand System Model

Congreso

Resumen *

The demand for alcohol is influenced by many different factors. Price and income arethe most important; however, alcohol consumption is also influenced by taxation, advertisingrestrictions, minimum age requirements, health concerns and household demographic andsocioeconomic factors like education, marital status, household composition and so on.There has been a few studies in few studies treating the demand for alcoholicbeverages in Argentina as a complete system and the estimation of expenditures and priceelasticities using a complete system of demand.The objective of this works is to estimate expenditure and price elasticity in differentbeverages purchased by household in Argentina. Beverages include: wines, beer, spirits, andothers. The method used to estimate elasticity?s is a system of demand using the Almost IdealDemand System proposed by Deaton and Muellbauer (1980). A two-stage estimationprocedure proposed by Shonkwiler and Yen is used . Data are draw from the NationalExpenditure Household Survey 2004-05, with a total of 21080 observations.Results show that expenditure has a significant influence in alcohol consumption andthe expenditure elasticities in wines are close to the unit, meaning that an increase of 1% inexpenditure, increase also 1% in quantity. Beer has expenditure elasticity higher than wine,1.13 and spirits show and elasticity lower than 1. Sparkling wine has a negative expenditureelasticity, having wrong sign, and one reason could be that only 0,2% of the household reporta positive compunction of this alcohol beverage in the sample used.The own-price elasticities are significant and with the expected sign. The own-priceelasticity in low-quality wine is more inelastic than in high-quality wine. Beer has own-priceelasticity less than the unity. In general, the alcoholic beverages studied respond more to totalexpenditure than prices.Demographic variables included in the model show an important impact in thedemand by household. Información suministrada por el agente en SIGEVA