Estimation of own and cross price elasticities of alcohol demand in the UK – a pseudo-panel approach using the Living Costing and Food Survey 2001 to 2009

Y. Meng, A. Brennan, R. Purshouse, D. Hill-McManus, C. Angus, J. Holmes, P.S. Meier

ABSTRACT

Background: The estimation of own- and cross-price elasticities of alcohol demand for different beverages and trade types (off-trade vs. on-trade) is valuable for the appraisal of price-based policy interventions such as minimum unit pricing and taxation. A few studies have attempted to estimate cross-price elasticities using either national aggregate time series or cross-sectional data, while the ideal dataset would be longitudinal panel data which are normally unavailable. A pseudo-panel approach offers the opportunity to apply panel models using repeated cross-sectional data. This paper aims to apply this approach to the UK Living Cost and Food Survey (LCF) 2001/2-2009 to estimate the own- and cross-price elasticities of off- and on-trade beer, cider, wine, spirits and ready-to-drinks (RTDs) in the UK.
Methods: A pseudo-panel with 72 subgroups defined by birth year, gender and socioeconomic status are constructed. Elasticities are estimated using fixed effects, random effects and standard ordinary least squares models with purchase quantity as the dependent, and price of alcohol and income as main independent variables. Additional independent variables are tested such as rates of unemployment and marriage and smoking status. Extensive sensitivity analyses are performed to test the robustness of model results.
Results: All estimated own-price elasticities are negative and 8 out of 10 are statically significant (p<0.05). Off-trade cider and beer are most elastic (-1.27 and -0.98) and off-trade spirits and on-trade RTDs are least elastic (-0.08 and -0.19). Estimated cross-price elasticities are smaller in magnitude with a mix of positive and negative signs. Model estimates are reasonably robust to different modelling methods and selection criteria for the pseudo-panel subgroups. Fixed effects models appear to be more theoretically plausible than random effects models.
Conclusions: The pseudo-panel approach offers a method to estimating price elasticities of alcohol demand using repeated cross-sectional data. Resulting estimates of own- and cross-price elasticities appear theoretically plausible and robust and could be used for appraising price-based interventions in the UK.

This paper has now been published