Beyond the mean: Estimating consumer demand systems in the tails

Abstract

Th e study proposes a novel approach to estimate price demand elasticities at the various levels of expenditures. Th rough the expectile estimator, the demand system can be estimated not only at the mean, as is generally done when implementing the OLS, but also at the lower and higher levels of expenditure. A simple demand system of equations focusing on fi ve basic goods: Food, Recreation, Clothing, Transport, Rent, using the Canadian Family Expenditure Survey data was estimated. Th e comparison of the estimated elasticity of each commodity at the selected expectiles is tested to verify the statistical relevance of any diff erence among the estimates in the tails and those computed at the centre of the conditional distribution. Th e results show with a strong evidence that the elasticity of Food and Recreation grows across the expectiles. Clothing variations are less evident, while the Transport and Rent elasticity is basically constant across the expectiles.