Survey Paper on Estimating Consumer Demand Systems

If you are looking for a nice survey paper on estimating static consumer demand systems (I was) Apostolos Serletis and William Barnett put one out a couple of years ago in the Journal of Econometrics. It's a nicely organized paper that should be understandable to anyone who's done the basic graduate level micro-economics and econometrics courses. In other words, it is really approachable compared to most papers published in the Journal of Econometrics :)

The beginning of the article reviews the basic neoclassical consumer demand theory. Following that, parametric approaches to estimating demand systems such as the Rotterdam model, flexible functional forms such as translog and AIDS, and "semi-non-parametric" forms such as the Fourier approach are covered. Next on the agenda are sections on revealed preference and on Engle curves. The final sections cover estimation issues, theoretical regularity - to what degree estimated demand functions meet the restrictions of neoclassical theory - and econometric regularity - mainly a discussion of non-stationarity - i.e. data with stochastic trends.

The only additional issue I would want the paper to cover, of course, is the question of how to interpret the estimated elasticities - are they estimates of short-run or of long-run elasticities - and how reliable the estimates from any single study are. The former is affected largely by the type (time series, cross section etc.) and properties (stationary, non-stationary etc.) of the data and the latter by sample size as I discuss for industrial interfuel substitution elasticities in my forthcoming paper and in my recent paper on estimating the emissions-income elasticity.