Regional Differences in Food Consumption in Urban Mozambique: A Censored Demand System Approach

A nationwide household survey for Mozambique is used to estimate a large censored food demand system with 12 food groups for the sample of urban households. Using the translog indirect utility approach, the censored nature of the data is addressed by estimating a system of Tobit equations with a recently suggested quasi maximum likelihood estimator. Augmenting the system with demographic and geographical variables in a theoretically consistent way, I find that differences in elasticities between regions are significant. The results show that regional variation has to be taken into account when evaluating policies or employing CGE models. Further, the approach employed here can be applied to a number of developing countries with varying geographic conditions.