Policy analysis frequently requires estimates of aggregate (or mean) consumer elasticities. However, estimates are often made incorrectly, based on elasticity calculations at mean income. We provide in this paper an overall integrated analytical framework that encompasses these biases and others. We then use empirically derived parameter estimates to simulate and quantify the full range of biases. We do that for alternative income distributions and four different demand models. The biases can be quite large; they generally grow as the degree of income inequality rises, the underlying expenditure elasticity differs from one, and the rank of the model increases.