Description

This paper explains the trad e-off between bias and variance in the choice of expansion factors for the Chilean Financial Household Survey (EFH). I outline several alternatives for the expansion factors assigned to each household in the EFH. The alternatives are based on a full post-stratification procedure using as st rata different groups of Chile’s geographical regions, the wealth of each to wn, and the income level of each household. I find that expansion factors based on a small number of strata can accurately represent the age, education and income distri bution of Chile with little bias and variance involved.